| JBL PB12 subwoofer, Class D amp, dead. - Click HERE for Original Thread |
| pat allen |
One of my friend just gave me a blown PB12, there is a class D module amp, where everything seems to be in.
The circuit seems to be repaired and looks like there was a lot of rework from factory. I know this unit never been opened by a redneck or so, my friend dont even have screwdrivers.
So, is there some schematics that i could get, or some inputs. I do think that the source of the problem is the module, wich is a small heat sink and some urethane molded behind it. When i turn it on, the woofer pops in with huge hummmmmm. Dead for sure.
Sources for this part?
It seems that there is a lot of bad coments on this subwoofer.
?
Thanks
Pat Allen |
|
|
| classd4sure |
| hmmmmmm Big pop and big hum huh? Check for blown output devices, sounds like they're toasted. Why they're toasted, I do not know. |
|
|
| pat allen |
There is no output devices, it is a monolithic device, and is a class D amp.
So, i f i want ot repair it, i have to change the whole unit (sealed)
What fears me is all the mess around it, wich dont seems good.
So i want to see a valid schematic before spending money in it.
Pat Allen |
|
|
| classd4sure |
Good news and bad news:
I've done some searching about this thing, I see what you mean about other users, basically 99.99893719% of everything I've read about it are users saying "it blew, loud hum, can't fix, known problem, faulty part"
.. ouch.
The rest of the % of what I read on it were of course manufacturers sales adds.
I've hunted down the likely name of your class d wannabe module "S64AMI power module" made by harman karding I think.
http://www.harmanaudio.com/search_b...BL&MarketId=HOM
It looks to me like they're only willing to sell you a replacement woofer.
It looks like this is certainly a faulty amp that should not be on the market, and like they're unwilling to replace it with anything that does work. It looks to me like they're still selling this scrap, and not fixing it.
It looks to me like I'll never be dealing with them!!
Based on all that, if the box /woofer itself is something you want to make use of, I think you can get it working by ripping out the electronics that are in there and replacing it with a good class d module, or plate amp.
Out of what is in there, you may only want to keep the cross over, and replace everything else with decent products, because I really don't think even if you could find a working S64AMI power module to replace it with, that you'd want to spend the time money and effort when it seems it will only blow again.
BTW, most users said it blew while idle playing no music, this is when most class d amps are switching fastest/more heat, maybe being in epoxy doesn't suit them well.
Hope that's more helpfull.
Chris |
|
|
| pat allen |
Wow ! that is a lot of information !
I did try to deal with the company, they offered me another module under some waranty and some fees, but i did ask to de guy how long will it survive, since they all blow.
I didnt get response from the guy then. Guess why.
The woofer in it is a 12inch one, big magnet, but a 2 ohm single voice coil. There is an electronic cross over in it (op amps TLO64), as well as the muting circuit.
The power supply is a very high voltage type, with a very small transfo wich is normal for a D amp. So, if i want to use any of its inards, i am stuck with a lot of issues, especialy the 2 ohm sub and the very small power supply.What is a plate amp?
The cabinet is in top notch condition.
Maybe it will find some use...for now it resides along with all my scrap amps and speakers.
thanks a lot for the info !!
Pat Allen |
|
|
| classd4sure |
You know, there's other "unhappy customer" alternatives, if being a nice guy about it fails.
Take it into the store with you, (hope your buddy kept the receipt) during peak time when it's full of people (saturday afternoon?) and scream your head off until they give you something new of equal value :)
Orrrr, get the email address of whoever is in charge at JBL, mention the vast amount of unhappy users you've found all over the internet, and the possibility of a class action lawsuit if the situation isnt' rectified.
Yeh, it's low and dirty, but so is selling garbage equipment for over a thousand dollars and not supporting it while hanging their customers out to dry.
The small transformer is not normal for a class d amp in itself, it is normal for a switching power supply, which as far as I know, is not the optimal set up for any amp as far as high end audio is concerned. It is a way for them to make alot of them very cheaply and maximise profits while quality suffers.
A plate amp, is just an amp on a plate :) Used for making powered speakers, cut the hole and screw them into the back of the speaker, pretty easy.
Go to www.hypex.nl and check out their DS1.2, run your mouse over the picture of it and you'll see it has a real power supply attatched, can hardly go wrong, it is also using cutting edge class d circuitry (so good it's almost a waste using it just for a sub), and the guy stands behind his products.
Shop around though and compare it to others you can find.
Anyway, I wouldn't worry about keeping their wimpy switching supply in it either, it's probably junk too.
Regards |
|
|
| JohnW |
Pat Allen,
I consulted for a company once that had a connection with JBL / HK - there Class D amps (nothing to do with me) where based on the Harris solution :(
Can you post a photo of the module - I have some spare modules here that MIGHT be the same.
John |
|
|
| pat allen |
Here is the complete board.
I meant by a small tranfo that is it small for a 250w amp, even D mode.
Pat Allen |
|
|
| classd4sure |
Hi,
I did more reading on these,
Either uses the S64AMI and S53AMI output modules, (both had problems) made by Analog Modules, Inc (the AMI in the serial #) they've been discontinued and JBL "was" making accommodation deals.
John, nice offer but considering all the problems people have had with these.. would you really want to replace it with a like unit, unless you wanted to fix it and sell it to someone you didn't like right away.
So is it a flaw in the harris design, if so, what, or the fact they encased the whole thing in epoxy?
Curious about this now :)
Thanks| quote: | | I meant by a small tranfo that is it small for a 250w amp, even D mode |
Got ya. |
|
|
| pat allen |
here the other side...
Pat Allen |
|
|
| JohnW |
Hi Pat (& Chris),
I'm sorry its not the same module as I have - could still send you one - if you want to try and shoe-horn one onto the plate - seem to recall they operate on single HBridge rail and -/+ 15 or similar for Opamp / modulator. However not sure if its worth it, as the modules I have are very poor performing - might be able to dig up measurements - if I can find them..
John |
|
|
| ssanmor |
If you want the schematics of the JBL PB12, please e-mail me.
Best regards |
|
|
| pat allen |
Ok, so here i am, i have opened the sealed module :) with the help of a milling. I have access to the power mosfets, wich are totaly ok.
ssanmor sent me the schematic of it all, the the inards of the module dont match with the schematic.
I think i will manage to build a new module myself, pcb,... |
|
|
| classd4sure |
Hi,
Nice work slicing that open.
If not the mosfets, can you tell us what is cooked on it? I read it is due to a flaw in the circuit, so I trust you plan on changing a few things.. where did they screw up with it?
Thanks |
|
|
| pat allen |
| I dont know yet, i havent removed all the epoxy, wich is difficult to not damage physicaly the components, i will surely fond what happened, and fix or improve it. |
|
|
| classd4sure |
| Neat, good luck, please keep us posted :) |
|
|
| pat allen |
So, i have mastered how to remove that stuff, man that stink !!
With a light flame torch, just heat it up without scraping it with a screwdriver too soon as i did, and it removes like a champ.
I went a litle bit too strong, but if i have another module i will be able to open it without messing it all like i did.
Anyway, i will do a full reverse on it.
the 2 chips are schmitt triggers, MC74ACT14AN, the small one is a LF358. There is also a small transistor, mpsa42.
All the others components are regular resistors, diodes and caps.
There is no SMT parts on this board. There is no reason why they did encapsule this part in epoxy. Is there a patent on this?
anyway, if i could only find the real schematic of it, i will figure what went wrong. Looks to me like one of the schmitt trigger have failed, since the power mosfets are ok.
any toughts? |
|
|
| pat allen |
| The under of the module, where i discovered to my absolute dismay that there was no components. :eek: :eek: |
|
|
| JohnW |
Hi Pat,
37 Minutes left of this year!
Are you sure one of the 14 Pin DIL is not a MOSFET Driver?
John |
|
|
| pat allen |
ha! i am in CANADA so it is 5:24pm here, sorry :D
No, they are realy schmitt triggers, i was able to read very clearly all the parts numbers.
The schematic that i have shows effectively some sort of driver, wich is a IR2111
IR2111 data sheet
I guess the module i have is the "old style" where they managed to drive the mosfet with the schmitt triggers, and the "better module" have the IR2111. It does not match at all with the schematic anyway.
pop me a message when you will celebrate !!! :grouphug: |
|
|
| JohnW |
Well at least from central Europe - A Happy New Year
Looking at the PCB tracking - it would apear that the MOSFETS are all Nch. If there are no MOSFET drivers - then I wonder how they drive them with the 74AHC14?
Circuit would be interesting...
John |
|
|
| pat allen |
Yes, they are both n channel mosfets. I am reversing the design now and hand write the schematic.
The v- mosfet is drived by one of the two schmit trigger chip, and is referenced to v- , no to ground, while the v+ mosfet is drived by the other chip, referenced to ground then.
Maybe one of the flaw in the design?
Happy new year John ! |
|
|
| classd4sure |
Happy new years guys.
Have a look at the supply line for the upper schmitt IC "driver" see if it's powered by a bootstrap cap/diode circuit. One thing I read about as well is caps going bad on these... maybe that's the one? Seems off for that to go but.... hey *shrug*. Sounds like reaaally weak drivers anyway.
I'm also surprised it's only a half bridge, if it is harris, all their patents I've seen seemed to use a full bridge. Maybe you can take the IC's off and test them in a dumby circuit for basic operation or something.
Cheers |
|
|
| pat allen |
Ok, here is my first drawing, duno if it is 100% acurate. It seems that there is a boostrap to power up the upper 74, with c5 and the diode on its right.
maybe i have done mistake while reversing it, but it seems quite simple.
Any tought???
thanks and a happy new year to every one !!! |
|
|
| pat allen |
Ok, after some testing on my bredboard, the schmitt trigger chip for the v- mosfet is shot, toasted, dead. :dead: All other components test good, even if they look burned (because of the way i have removed the epoxy, with a flame torch :hot: ).
Looking at the schematic, there is not protection diode on its vcc line, while the upper one, got one. You should see a zener diode at the right of c5 rather than a plain diode.
The vcc for this is issued from the bridge rectified -95v, with some diodes and 3x 10k resistor in parralel, making a -89v if measured from the ground. This chip (IC2) is referenced to the -95v, so it got +6v at its vcc.
I will build a prototype with better voltage limitations/protections for both chips, with more caps for filtering, there is very few of them, actualy only 0,1uf.
The opamp is also very badly powered. It is by the means of 3x 4.7kohm resistor in parralel, 2w, and a 15 volt zener diode, same circuit for the +vcc and -vcc, making a poors man +15 0 -15. Source is directly from the bridge rectified +95v and -95v :hot: There is no decent caps on this too, only a 100nf on the main board of the whole unit. what a sh*t.:dead:
Sorry for my poor schematic, i will do a better one with my cad (power logic) when i will go back to work. |
|
|
| classd4sure |
Hi, nicely done I'm impressed!
I think I see the basics on how it's trying to work. Simple for sure.
Few things I'm unsure of.
After the bootstrap circuit, there's a diode, a 5kR, then an inverter, the positive supply of that inverter looks like it's connecting to to all that, is that the case? AAAAhhh, the 5kR must be a collector load. I bet that diode must be a 12V zener.
I'm not seeing anything that would charge that bootstrap cap on power up.. but I see you have a bit of a line/something drawn up above that doesn't seem to go anywhere, are we missing anything from this?
What connects to that... inverter? that goes to the BJT's emiter? Not the feedback point?.... I can't figure that part out at all, but I don't see any other place for a feedback input, John? :)
IRF640's... switching pigs. You can certainly improve on that choice.
I have a feeling there's an error in those "driver" voltages.
Being N channels you have to bring gate +~12 with respect to source to switch on, it looks like the ground is wired to -95+12, it would always be on and couldn't turn off like that.
All the rest seems like it could make sense to me :)
Regards and Happy New Year |
|
|
| classd4sure |
You were posting while I was writing.
Nice work you're on the ball! Don't need my 2cents but I cant' stop myself, nothing on TV.
So what do you suppose that inverter off the BJT is for? I can take a few guesses but.... nah, not at this hour.
I knew something was fishy with those voltages. 6V was my first guess but it didn't make sense so I kept looking at it, hard to make out the writing too. I would think that would have burnt the mosfet out before the inverter though, not turning it fully on. If the driver is burnt, the mosfet must be, are you positive it's OK? Burnt mosfet would go along way to explain that loud hum you had too. Wow, how'd they let it go out the door like that? Something to be said for DIY alright.
There needs to be some type of means to charge that bootstrap cap on power up too, with a delay, you already have the zener there, you just need two more resistors, one from positive rail to one side of the zener and one from the other side of the zener to the negative rail (thanks John), maybe you have a delay with it already too? You dont' really need another zener on the lower one, not if say it were comming off a 12V linear regulator IC.. or any "decent" regulated supply of some sort, which it sounds like you'll be adding.
Maybe since you're redoing it you'll consider using a decent gate driver IC, but if that worked with those mosfets, adding decent mosfets for switching should be enough of an improvement to let you sleep at night.
This was fun, thanks! |
|
|
| classd4sure |
:blush: funny, yet, not really.
Right out of no where, dawns the light.
That inverter on the bjt+diode... start delay / precharge. Seems you knew that "SD" Start delay.. You know what that's the coolest part of the whole amp!
Anyway, you seriously did an excellent job reversing this thing, and in no time at all.
So, is there no feedback on this thing? |
|
|
| JohnW |
Hi Pat & Chris,
Well done in tracing the circuit so quickly.
The modulator circuit doesn’t make much sense – where’s the oscillation element? I would have expected to at least see a feedback path from the OPS to the first integrator via the integrator Cap or a resistor… If the Cap was connected from the OPS then it would look a little like UCD / hysteresis switcher…
A Single ended Sub amp – now that’s just asking for trouble! - now I wonder why the PSU caps are so big ;)
John |
|
|
| pat allen |
Ok, i have tripple cheked the boostrap circuit, i did a litle mistake, the zener is on the right side of the 5k resistor, not directly in parallel with c5.
D6 might be responsible of some sorf of initial start?
The RF connector on the top left corner is connected to nothing on the pcb, but is there.
i am not that much familiar with class D amps, but i see the main thinking behind it all.
What surprises me is the total absence of active protection, wathever is happening, full dc voltage will go thru the woofer.
Yes, i am 100% sure of that, both mosfets are ok.
On the main board, there is a feedback resistor from input to output. 56.2k.
SD is some sort of protection where it stops the unit to oscillate, but isnt there somethin wrong with this, one of the 2 mosfet will stay on then?
On the other schematic i have (original PB12 power amp module version 2), they show a lot more filtration caps for the opamp, and the IR2111 driver, so they knew it was a flaw in the design .
There is a relay and a temperature sensing on the "better" circuit. Maybe i will inspire myself in doing my protoype :D
Thanks for all the good comments. |
|
|
| JohnW |
Hi Pat / Chris,
With the Feedback resister connected to the integrator input, the modulator circuit now makes more sense. In my opinion, the circuit is a Hysteresis “UCD” type modulator that pre-dates the UCD patent - if the AMI 1999 date is correct.
John |
|
|
| pat allen |
Hi, do you have info about the UCD patent, or some links that could explain to idea in the design?
thanks a lot. |
|
|
| JohnW |
Hi Pat,
Take a look at Chris thread "UCD Based Fully Differential Full Bridge 450W RMS Amp"
Sorry dont - know how to post thread-links.
John |
|
|
| classd4sure |
Good morning,
MM.. coffee good.
Ok, D6 that's the bootstrap diode? Yeah it kind of is responsible for precharge in the sense that it is in the precharge path.... if I'm right about this that is.
I'm not 100% sure on this, but on power up SD is probably brought high, which brings node/pin 8 low on the inverter, creating enough of a drop to forward bias the diode, turning on the BJT and pulling down to the negative rail, once that is done it charges up the bootstrap cap to... 6volts :dead: through D6.
When SD is brought low, node 8 goes high, forces the diode into blocking mode /disabling the precharge circuit and then the bootstrap cap is ready to "fly" all charged up.
Standby mode must just hold SD high keeping it in precharge mode as well. The only way I can see it keeping the lower driver disabled is by syphoning off current through the base to prevent the lower driver from being driven, maybe this explains the use use of that 1K resistor in the lower driver path, limit the current to the base so this balancing act can work. That's just alot of speculating though, thoughts?
So I guess that 5K is doubling as both a current limiter for the zener and acting as a pull up for the collector.
Can you find out what the zener voltage is? At this point I'm getting curious as to wether or not they used 6volts drive voltage intentionally to help tame the switching behavior of the mosfets, I'm shocked that they weren't burnt, impressive, I still think they're pigs though, they have a fairly high Ron too.
Thanks |
|
|
| Tekko |
| That amp is modified by a owner for sure, that would explain the big ps caps, about the oscillator element, it could be self oscillating or the oscillator was on the main board. With a voltage at 95 volts per rail the current is so low that the transformer doesent need to be any larger. |
|
|
| classd4sure |
Hi,
I doubt it was modified, it was encased, Pat had to cut open the case and chisel out the epoxy. There seems to be have been several revisions of it at factory + the amp module changed entirely though, so there's alot of variations of it floating out there I'm sure. (Both amp module versions eventually were cancelled due to the all the problems they've had with them).
The big PSU caps, well, it is a half bridge sub woofer, small caps at low frequency on a half bridge would pump up pretty good I think. |
|
|
| pat allen |
It havent been modified, the big caps were added by the company that did service it first time, it is part of a recal in the service manual.
SD seems to be always high, it is pumped externaly by a zener diode (1n5265b, 6.8v) and a 8.2mohm resistor, making some soft start because of the high r value, and the C11 that could be around 0,47uf. The short circuit protection that is on the main board will put SD low (-95) if activated :dead: there is 3x 0,1ohm resistor on the ground path of the speaker output, wich will activate a pair of transistor to activate another one, and then put SD low, in case of a short at the output, maybe it limits the overall power output?
I am slowly understanding every components, i have also measure some caps and resistors to complete the schematic, the only 2 missing are the C1 cap on the first op amp, and the cap above the 1k resistor marked with a ? C9 is 47p. |
|
|
| Tekko |
| Okok, good luck with the searching for what made it blow :) I got a class D amp myself, its my own brand but based on the B&O ICEPower 250A module. |
|
|
| classd4sure |
So they must have been suffering from supply pumping and blowing up supply caps. That's consistent with some posts I've read too about blown supply caps, and they just fixed it by adding bigger caps, which is fine.
Is there a chance that circuitry on the output is just there to sense current flowing, so that it knows it is oscillating and disables the pre charge circuit? You don't need it once it is oscillating as it refreshes the cap with each cycle.
I just dont' see how that bjt can disable this thing. I'd have to simulate it, but thankfully, that's your job:) |
|
|
| classd4sure |
I was going to simulate this thing, but.... :xeye: Maybe I will :confused:
Anyway, no wonder it smoked, can't believe it ever worked at all, whoever "designed" this was dancing on the razors edge blindfolded and loving it.
Your schmitt trigger hex inverter, you said was MC74ACT14AN, don't know what that last "A" is for.. , it is part of the "ACT" cmos family, advanced cmos, TTL compatible inputs.
Characteristics of this family, Vcc MAX=5.5!
They were running with >6, and unregulated.
Also, you're just not switching an IRF640 fully on with 6volts gate drive, they must have run hot.
Options:
You could use a 10V cmos inverter, but that would be much slower (1 or 2Mhz as opposed to the 100Mhz the ACT family is good for) and the threshold will be alot higher, around 5 volts I think, might not work very well.
Or, could possibly add some type of follower or other gate driver, extra parts, extra supplies, complication..yuck.
My idea: rebuild as is, but fix the supplies like you planned, regulated to 5.1 volts, and use a logic level mosfet.
Check this one out:
At Vgs=10, Rdson=.075 Ohms
At Vgs=4.5, Rdson=.08 Ohms
Vth min = 1volt
Vth max=2volts
http://www.fairchildsemi.com/pf/FQ/FQA34N20L.html
I think that'd be a big improvement, with no added complication or significant redesign. |
|
|
| pat allen |
where do you see a bjt? there is only a regular NPN transistor (MPSa42).
The initial switching frequency is 100khz, as the sevice manual says.
I have the data sheet of the MC74ACT14AN and the vcc is 7v max.
Yes, having some sort of ttl could bring the voltage to a safer condition for the mosfets, i will try all this anyway with fresh parts.
Do you have any suggestions for a mosfet?
Is the IRFZ4x serie worth a try?
I got plenty of IRFZ44.
There is realy something wrong with the schematic for the SD input, there is a zener diode tied to v+, a 33k to it tied to v-, and the 8.2mohm resistor is connected to the tied of the zener and the resistor so on should read there +vcc-Vzener, while it should have -vcc+Vzener to get around 6.2v to the SD input, referenced to -vcc. :dead:
If i take measurements on the circuit while the module is not there, i read 88.2 volts.:bawling:
The zener and the resistor just needs to be swaped...:confused:
There is UVLO writen around this part of the schematic.
thanks. |
|
|
| pat allen |
Someone will have to explain me something:
look at the schematic.
This is what the original schematic shows. the zener diode is a 65volt one.
SD input of the AMI module should never be more than 7 volts, referenced to -95v.
I am getting 6.2 volts with the schematic is show you now, so it is "ok"
i dont get it, i was expecting to see +95-65(Vzener)=30v, referenced to ground, but something like 125 volts if referenced to -95v:confused:
The 8.2mohm resistor is responsible to charge the capacitor at the input of the schmitt trigger inside the AMI module, as a "soft start" ? |
|
|
| pat allen |
I was expecting something like this.
The Vmax sent to SD cant be now more that the Vzener.
:confused: :confused: |
|
|
| pat allen |
Sorry for my trillion post in series, here is the main board schematic.
Comments please !!:apathic: |
|
|
| classd4sure |
| quote: | Originally posted by pat allen
where do you see a bjt? there is only a regular NPN transistor (MPSa42).
The initial switching frequency is 100khz, as the sevice manual says.
I have the data sheet of the MC74ACT14AN and the vcc is 7v max.
Yes, having some sort of ttl could bring the voltage to a safer condition for the mosfets, i will try all this anyway with fresh parts.
Do you have any suggestions for a mosfet?
Is the IRFZ4x serie worth a try?
I got plenty of IRFZ44.
There is realy something wrong with the schematic for the SD input, there is a zener diode tied to v+, a 33k to it tied to v-, and the 8.2mohm resistor is connected to the tied of the zener and the resistor so on should read there +vcc-Vzener, while it should have -vcc+Vzener to get around 6.2v to the SD input, referenced to -vcc. :dead:
If i take measurements on the circuit while the module is not there, i read 88.2 volts.:bawling:
The zener and the resistor just needs to be swaped...:confused:
There is UVLO writen around this part of the schematic.
thanks. |
Hi, The MPSA42 NPN is a BJT.
I checked the data sheet again, it does say VCC MAX =7Vdc. But that's the smoke point! "Maximum ratings beyond which damage to the device may occure"
While the "Recommended MAX values" are VCC= 5.5Vdc.
Even if you wanted to push your luck with the chips and run them close to 7 volts VCC, it still isn't really enough to drive a regular mosfet as a switch comfortably, and if you dont' turn them fully on they'll run hot and might burn.
The speed of the 10V cmos family would likely be just fine at ten times more than the switching frequency, but they still have a much higher threshold which might complicate things, or degrade performance, so I would stick with the ACT family and use a regulated 5Volts for VCC.
You can not use an IRFZ44 with this, It is only a 60V part, it isn't a logic level mosfet either (Vth is 2 to 4 volts), it is only starting to think about turning at that level of gate drive, and needs to be driven up to at least 8 min, 10 comfortably, to get the lowest on resistance.
Have a look at the link I gave you in my last post, it is a 200V part like the current device is, which you need because with +-95 volt rails, they have to block 2*95, with some safety margin preferably), and handles more current than the mosfets that were in there. They will be fully turned on with 5Volts gate drive, they seem a much better choice than what was in there.
The beauty of this is, if you use those mosfets, you don't need to redesign anything, can still use the same PCB, with all the other parts staying the same!! Just swap mosfets, replace all components that got burnt, and improve the supplies like you were going to (regulated to ~5.1 not 7). I'd find some way to heatsink the mosfets before you test it too, the epoxy was doing that job before, if you fire it up with no cooling they'll smoke pretty quick.
I figure that'll work fine unless some other problem we haven't spotted yet creeps up on you.
I really can't make out that whole SD thing, maybe someone else can help with that.
Regards |
|
|
| pat allen |
Hi, you are right, i didnt look deep enough about the bjt, my mistake, and for the irfz, i didnt rememeber that they were low voltage, now i know why i have so much of them, i have used them for a pwm power supply for a car amp.
I never meant to drive the 74act beyond theirs limits, and it is one of my goal to "protect" them. Building a 5 volt regulator could be very cheap and effective, if i can find those FQA mosfets, i usualy buy my parts at Digikey, they dont have it :(
I will build some prototype tomorrow, with all the mods i plan to do.
That will be fun, its my first digital amp project !
One last question if you dont mind, i am wondering why the psu is so high in voltage, the current capability of the tranfo seems to be very small, the 95v will never to all thru the sub. They claims a 250w rms for this set-up.
Could it be smart to lower the voltage and raising up the amps?
I have looked at the ZaP pulse modules, those modules needs a lot of amp to deliver the advertised wattage, with far lower voltage...could it be only due to %efficiency ?
JBL=65%
Zap=+95%
I am just wondering where i can improve :D ...
thanks for all the good information you give me. |
|
|
| classd4sure |
You can improve by ripping it all out and buying that plate amp I linked you to awhile back :)
I have no idea how they calculated that power rating, but expect it to be 90% BS. By my math the woofer would need to be over 250Ohms to get around 250W RMS sine power.
Know what the impedance of it actually is?
Anyway I'll spend a few coffee's thinking about that SD thing, I still think it is just a pre charge control circuit, if it puts the oscillator in stand by, there's some voodoo happening.
Oh, you might be able to get some samples of that mosfet direct from fairchild. If you fish through what Digikey has you may find other "logic level" mosfets, but they might be rare, I haven't seen too many logic level power fets, fairchild make really good mosfets anyway too.
Got any ideas how you can cool them while you test?
Regards |
|
|
| pat allen |
Hi, i finaly made this sh*t work on a breadboard.
The basics seems to work, i am able to get sounds out of it with 15 volts rails.
I have redone the schematic with power logic. There is still some parts that i dont got the value.
C3 and C4 are not on my proto and it seems to work good, i did try some caps in there, from 10p to 100n, no changes. Duno value.
I have a signal generator and a techtronics scope and can see the actual sine signal if it gets distorted or bad,...it is fairly good for now until cliping wich occcurs at close to vcc.
C6 also bugs me, it seems that it just limit the free bandwith of the osccilation, even if it oscillates at only ~20khz for now, cant make it go to the 100khz it should. C7 also seems confusing, i have replaced it with a resistor because the op amp was not able to lower its output enough to trigger the first shmitt trigger.
So C7 is a 4.7k for now. C2 also is an unknow cap value, i have put a 0,01 for now and it works.
SD when high ground the emitter of Q3 thru d8, enabling the circuit to self oscillate. It is part of the muting circuit and soft shut down, it enable the amp to stop without a thump.
the upper shmitt trigger is powered thru D6, when the signal input is low, it charges c1.
the overall circuit is also very instable and very dependant of the output filtering stage. as soon as i put inductor/caps, the free oscillating frequency drops a lot.
Let me know what you think of the missing value components. |
|
|
| Tekko |
| **** what a simple Class D amp :bigeyes: My class d am has alot more components, but my amp is intended for fullrange tho. |
|
|
| pat allen |
Please, take note that the schematic is the AMI module only, there is several other components on the circuit, see one of the post above for the other schematic.
It work full range for now. |
|
|
| pat allen |
Yes, but no close to the quality of yours, you work for Bang Olufsen ? :bigeyes:
This is still something JBL did sold, again, :bigeyes:
Duno how this sub cost new...
I have changed the power supply on my bread board from 12v rail to 30 v rail, and the mosfets are irf840 wich are not very suited for the task. There is no heat sink as well on them:hot:
I managed to get 25v pk to pk on a 4 ohm load, with some begining of audible cliping.
I am stumped. It smell bad too, the mosfet are glow red :D .
This is roughly 75wrms i think.
I will order some better mosfets and will build a pcb, there is some emi that i can see on my scope.
This begin to be a fun project, i have tweaked the caps and now i got around 100khz free running frequency.
It does work :eek:
cheers !!! |
|
|
| classd4sure |
Hi, what are you final cap values?
I would expect the mosfets to glow red and burn out very fast. This is why we want to use logic level mosfets, so they turn fully on with such low drive voltage. Otherwise they'll be passing full current with a high Ron and heat like crazy. Some of your values might need to change once you put them on a PCB because it will have less stray junk to worry about. You could try doing a dual layer PCB and use the top layer as a dedicated ground plane, would help with the noise.
Tekko, are you sharing your design with us?? It's not like you can sell it anyway.. :devilr: |
|
|
| pat allen |
Tekko, are you sharing your design with us?? It's not like you can sell it anyway..
good one !!!:D
Surpisingly, i can play music quite loud and they dont burn...i dont care because i have tons of them, we use this mosfet here in one of our product so...
Final caps are: c6 is a 10p, c2 is a 0,047u, no c3 and c4.
There is a missing one at the emitter of q3, that goes to -95v, is a 0,1uf for now.
R7 is a 330 ohm.
C5 is a 0,1uf
R10 is a 1kohm, i found that r10 ajust the symetry of the wave, i get the best results with 1k, and a 0,01uf in parralel on it, it also removes some noise at very low frequency (5hz and less :D )
There is still something that i have not figured the why of, is the voltage that flows thru D6. In order to make this amp work, i have to send dedicated voltage thru it, so figure for now that D6 is a 1k resistor connected to +vcc. It work fine as is.
Feedback resistor is still a 56k.
Mosfet are now ordered (the ones you suggested me)
tanks. |
|
|
| Tekko |
| Pat, no i do not work for B&O. Why ? |
|
|
| pat allen |
you refered "simpler than my amp" and then given a link to BO amps...
:confused: |
|
|
| Tekko |
| I just meant that the B&O module has alot of ic´s and components on it. and your jbl thing only had like 2-3 ic´s that was only smitt triggers and a quad opamp ic. |
|
|
| classd4sure |
| quote: | Originally posted by pat allen
Tekko, are you sharing your design with us?? It's not like you can sell it anyway..
good one !!!:D
Surpisingly, i can play music quite loud and they dont burn...i dont care because i have tons of them, we use this mosfet here in one of our product so...
Final caps are: c6 is a 10p, c2 is a 0,047u, no c3 and c4.
There is a missing one at the emitter of q3, that goes to -95v, is a 0,1uf for now.
R7 is a 330 ohm.
C5 is a 0,1uf
R10 is a 1kohm, i found that r10 ajust the symetry of the wave, i get the best results with 1k, and a 0,01uf in parralel on it, it also removes some noise at very low frequency (5hz and less :D )
There is still something that i have not figured the why of, is the voltage that flows thru D6. In order to make this amp work, i have to send dedicated voltage thru it, so figure for now that D6 is a 1k resistor connected to +vcc. It work fine as is.
Feedback resistor is still a 56k.
Mosfet are now ordered (the ones you suggested me)
tanks. |
Yeh, as long as they dont' melt into your proto board and touch each other :hot:
Your earlier posted seemed to confirmed how that whole BJT mess seems to work to charge the bootstrap capacitor..
But it won't do it without an additional resistor?? hmm is the turn on delay working as it should?
It's OK to add an aditional resistor to VCC to the bootstrap cap, but I hope you didn't actually "replace" D6 diode for a resistor! If you're going to leave that resistor there it would be a good idea to add another resistor after the cap tied to the negative rail, and with enough of a delay you can make them bigger as well.
Hope those mosfets work well for you. |
|
|
| pat allen |
got a new thing...
It seems that the bandwith is very narrow. Like from 20hz to 80 hz.
I got a 250uh inductor coil with 2x 2.2uf caps as the only filtering devices at its output.
It feels like when i am getting the resonant frequency of the sub (8 inch subwoofer, 4 ohm from a infinity amplified subwoofer), that it got the best results. I can throw 22v pk to pk thru it without having distortion, and it occurs at 33hz. The mosfets dont get hot at all.
If i change my frequency generator up or down, then i can hear some cliping as if i was hearing the mosfets turning on and off.
Any ideas? it only occurs at high levels, the lower it is, the better the bandwith is.
I am looking at the bjt and if the signals are not overlaped.
thanks. |
|
|
| pat allen |
wana see pictures??
first one, the ~100khz operating frequency... |
|
|
| pat allen |
And, output signal with a 4 ohm load...
30.8 volts pk to pk...do the maths :bigeyes:
You can see some hf stuck to the signal, i guess it is because it is on a bread board...poor ground plane,... |
|
|
| pat allen |
And, how it looks on the bread board...
I know, it is ugly :D |
|
|
| pat allen |
and that poor litle subwoofer...
yes, it does wooof, at 35hz...
cant wait to build a board and the better mosfets...
I will do a full schematic upgrade, with all the component needed to make it work, and the gerbers for the PCB when it will be done.
I will do it a la zappulse i think...
cheers !!! |
|
|
| classd4sure |
Good stuff.
Looks like there's some high frequency ripple/oscillation on your square wave, the gate turn on resistors are pretty small though, could be the cause of that??
It would be really cool to see a before and after of that with the new mosfets when you get them, I dont' think you'll have to wait very long for those either btw.
I see you do have some means to cool the mosfets after all (better than nothing).
There probably isn't much point trying to tune it more than it already is on the breadboard, alot will change once you get it on PCB.
It would be cool of you if you took more pictures of how it worked when on the new PCB, and then again with the new mosfets.
You might find your biggest problem is keeping those new mosfets turned off when they're supposed to be, try to keep all the current loops from the driver to the fet as short as possible, ground the drivers right on the source pin if you can.
Thanks |
|
|
| pat allen |
I realy cant make that boostrap work, it only work if i deliberately apply vcc thru a resistor to the upper chip.
See how r17 replaces the d6 diode in the previous schematic.
D7 is a 5.8v zener. |
|
|
| Tekko |
| Impressive work, Pat :bigeyes: :) Almost so i wanna build a class D amp myself. |
|
|
| classd4sure |
Hi Pat,
If you wire the bootstrap exactly as you've drawn it in your first "reverse" schematic I think it would work fine.
The bootstrap circuit will only charge the cap when the amp switches low, so check and see if your SD circuit allows the cap to precharge before the output stage is enabled, if it does it should work fine. |
|
|
| pat allen |
Done this way, it just dont want to start it self osscilation, if i apply few seconds my frequency generator on it set at ~100khz, then it start to oscilate by itself. Unless than that, it stays stuck at -vdd.
:confused: |
|
|
| classd4sure |
| If that's the case I dont' think that SD/inverter circuit can be working to allow it to charge, maybe you can monitor the voltage across the bootstrap diode while you pull SD high/low and see if it does anything? |
|
|
| usekgb |
I hated this sub that JBL put out. I am a JBL service center, and this amplifier module has been nothing but a headache. Why would JBL not use a Crown (also a Harman International company) design in their powered speakers? The amps that were in these PB12's had more problems than just the amp module going dead. I've seen C6 explode several times, taking out the transformer and all surrounding circuitry with it. I've also seen traces lift off of the PC board near the output resistors. This same design and amplifier module was also used in several Infinity subwoofers. The jacks have failed, the PCB's have poor soldering jobs, etc...... Why can't JBL just make good quality products like they used to? IMHO, they are ruining their reputation as a quality loudspeaker manufacturer with these sub-par products that they keep putting out.
Cheers,
Zach |
|
|
| pat allen |
Well, the sd thing dont change a thing...thats it, it just feels like it is a shut down where it whould stop to make it oscillate, duno.
the free oscillation is very sensitive, if i modulate a litle bit too much on the input, in stops to oscillate and goes -vcc or a crazy 1khz +vcc, -vcc.
This is with the bootstrap as it should...the best it worked it was when i force feed to upper shmit trigger...
what a piece of ****.
I think my op amp is bad, at least, it is the op amp that was in the module :xeye:
Lets asume that it is bad. |
|
|
| classd4sure |
Hi Zach,
I think I can take a pretty good guess why...
Look at the circuit! Total build cost (mass produced).. I'm not sure.. less than a dollar???
The PSU caps exploding.... because they pinched more pennies and sized them too small.. Obviously, quality is not a concern for them. Just think, someone probably got a big fat check for recommending this topology with respect to every penny they pinched.
Kept you busy though, didn't it? I only hope this screw up cost them far more than it made them! I noticed your key words "that they keep putting out" ... :xeye: |
|
|
| classd4sure |
| quote: | Originally posted by pat allen
Well, the sd thing dont change a thing...thats it, it just feels like it is a shut down where it whould stop to make it oscillate, duno.
the free oscillation is very sensitive, if i modulate a litle bit too much on the input, in stops to oscillate and goes -vcc or a crazy 1khz +vcc, -vcc.
This is with the bootstrap as it should...the best it worked it was when i force feed to upper shmit trigger...
what a piece of ****.
I think my op amp is bad, at least, it is the op amp that was in the module :xeye:
Lets asume that it is bad. |
Yeah looking at it again I dont' think the SD circuit is capable of doing anything in that respect. There must be some associated circuitry on the other schematic to help this function, I am unable to follow it though.
Simple cure.
1. Add something like a 1k resistor + diode from +95 rail to the top of the bootstrap cap. You can use a zener there a bit smaller in value than your Vcc supply which sits on the negative rail uses, this way I think it will be disabled while operational, which is good.
2. Add another resistor from the ground side of the bootstrap cap (tried between the mosfets) and tie it to the negative power rail.
Should be pretty high in value, 50K?
You can get it working without the extra zener too, but it will always be taking some charge so use a big enough resistor, and just give it enough of a delay at turn on to allow it to fully charge.
See what that does for ya. |
|
|
| pat allen |
| quote: | Originally posted by usekgb
I hated this sub that JBL put out. I am a JBL service center, and this amplifier module has been nothing but a headache. Why would JBL not use a Crown (also a Harman International company) design in their powered speakers? The amps that were in these PB12's had more problems than just the amp module going dead. I've seen C6 explode several times, taking out the transformer and all surrounding circuitry with it. I've also seen traces lift off of the PC board near the output resistors. This same design and amplifier module was also used in several Infinity subwoofers. The jacks have failed, the PCB's have poor soldering jobs, etc...... Why can't JBL just make good quality products like they used to? IMHO, they are ruining their reputation as a quality loudspeaker manufacturer with these sub-par products that they keep putting out.
Cheers,
Zach |
Hi Zach, since you are a JBL service center (still?), you might have some good connection with the company?
What i am looking for is the original schematic of the S64AMI module, because i cant figure some components values in mine.
If you cant, do you have someone that i can reach and ask.
Thanks.
BTW, i melted everything.:dead:
It never worked as it should, the best i did get is with a lot of tweaks, and it was not that much good.:mad: |
|
|
| classd4sure |
| quote: | | Hi, what do you think of using half the S64AMI (op amp inputs and comparator right thru the first Schmitt trigger) with the half bridge driver of the previous schematic? |
May as well continue this here..
Did you make the PCB for it? I wouldn't expect it to be anything close to good until it's on a PCB, and you'll have to re-tweak it once it is, what you do on the protoboard is just good for a starting point. You can probably measure the value of the caps with a cheap multi meter right?
How come you cant' start it? Still having bootstrap problems?
While you're still fighting with this aspect of it I wouldnt' be so quick to make any radical changes to the design, but you can use a gate driver IC with it of course. Would it be simpler? You'll need at least another power supply, your old SD circuitry will be useless.
Also, while you can use a gate driver, I really wouldn't implement it as per that posted schematic, and it wont' simplify anything, you'll need extra support circuitry and an additional power supply for it, you'll gain the advantage of not needing logic level mosfets.
I believe they used those extra transistors simply to increase drive current, shouldn't be required, use a better mosfet, another reason behind that, they may have been trying to combat the crappy rise time of the IC, but it also say's it has an internal deadtime .... ~650ns!!! I wouldn't use that chip either.
Regards |
|
|
| Tekko |
| I´m impressed of your work, Pat :cool: |
|
|
| pat allen |
| quote: | Originally posted by Tekko
I´m impressed of your work, Pat :cool: |
Thank you, hi havent done anything that much incredible, maybe just opening the AMI module is something...
classd4sure : why do you say that i will have to redo the power supply if i want to use the IR2111?
-because of the high voltage rail and low current?
-because of the new requirement of a +15vdc referenced to the -vcc?
it already have a +15, 0,-15 for the op amp, and for the +15vdc referenced to the -vdc, there is already a +6vdc referenced to -vcc, all it needs is to change the zener value, this voltage was used on the AMI for the bootstrap, same thing with the IR2111.
The digital mosfets you suggested me are NA for now, at all our distributors, and even directly at Fairchild :(
I plan to order few of thoses Ir2111 anyway, and play with them.
thanks. |
|
|
| classd4sure |
Ok you're right, no need for new supplies.
I'd get the IR2101 instead though, less internal delay. |
|
|
| classd4sure |
Ok you're right, no need for new supplies.
I'd get the IR2101 instead though, less internal delay. |
|
|
| pat allen |
Hi, there seems to be a lot of IR driver from international rectifiers, i have looked at the IR2104, it have a shut down circuit where it shut off both outputs, could be good for the subwoofer application where when it is not used (waiting for music to get in), it could be not oscilating then no possibilities of burning it??
And what about the self oscillating ones?
One could modulate the frequency on its input and might get some sort of ultra cheap D amp???
I will order all of them and play :D |
|
|
| classd4sure |
Self oscillating ones?
We have to note the difference between controller and driver here.
Typically the controllers, while they look perfect for the job, are not meant to handle the kind of voltages audio requires, they usually lack in some respect.
I think you're right on about the disable pin. Also some drivers only have 1 input, others have 2. This is one difference between the ir2111 and ir2101. I would design the circuit fo the IR2101, for it has much less internal dead time, and you can gain full control over both inputs, it is a recommended part for audio. |
|
|
| pat allen |
Did you look at the app notes of IR?
there is an interseting and simple design for a 200w+200w with the IR 2011, wich looks a lot at what i want to achieve.
On the top of your head, do you have others suggestions of more commonly available power mosfets?
the IRFB23N15D seems to be used a lot in D amps, but feature a 0,1 ohm RDS on, wich i found high compared to others with 0,0xxx ohms rds on.
thanks. |
|
|
| classd4sure |
Simple ? :cannotbe:
I can't recommend any MOSFETS off hand..
I'd keep looking for a simpler design, with feedback taken after the filter for best performance. I admire your ambition! |
|
|
| cantech |
Dear Pat,
I have parts and modules for your specific amp. According to the picture your amp has been repaired without original parts. I have original parts for your amp. Approximate cost is 80 dollars CAD. If you are interested PM me for contact information I am based in Toronto, Ontario. |
|
|
| classd4sure |
| quote: | Originally posted by cantech
Dear Pat,
I have parts and modules for your specific amp. According to the picture your amp has been repaired without original parts. I have original parts for your amp. Approximate cost is 80 dollars CAD. If you are interested PM me for contact information I am based in Toronto, Ontario. |
You're really goign to sell him a part which is based on a defective design for 80$?? C'mon man, a UCD 180 is 100 Euro and it isn't defective. Give it away or throw it out :smash: |
|
|
| pat allen |
Thanks, i already answered him. I have ordered all the parts to build a IRAUDAMP a la International Rectifier, i got all the components to build 4 modules (1 chanel each) for around 50$ CDN...i have build my own PCB (1ch) for testing, i will build a better one with 2 channels for a bridged mode and throw a lot of watts, wich will be cool for my TL project for my home.
Thanks. |
|
|
| usekgb |
| quote: | Originally posted by classd4sure
Hi Zach,
I think I can take a pretty good guess why...
Look at the circuit! Total build cost (mass produced).. I'm not sure.. less than a dollar???
The PSU caps exploding.... because they pinched more pennies and sized them too small.. Obviously, quality is not a concern for them. Just think, someone probably got a big fat check for recommending this topology with respect to every penny they pinched.
Kept you busy though, didn't it? I only hope this screw up cost them far more than it made them! I noticed your key words "that they keep putting out" ... :xeye: |
Unfortunately, JBL never offered a fix for this problem. I actually spent more time on the phone getting replacement subs for customers than I did actually fixing the dead ones. JBL (and Infininty) ended up offering customers accomadation pricing for replacement subs that were still under warranty instead of fixing the problems that they had. And, I'm sure that this SNAFU ended up costing Harman International quite a bit of money over the course of this whole mess.
| quote: | Originally posted by pat allen
Hi Zach, since you are a JBL service center (still?), you might have some good connection with the company?
What i am looking for is the original schematic of the S64AMI module, because i cant figure some components values in mine.
If you cant, do you have someone that i can reach and ask.
Thanks.
BTW, i melted everything.:dead:
It never worked as it should, the best i did get is with a lot of tweaks, and it was not that much good.:mad: |
Here is a link to the PB12 service manual:
PB12 Service Manual
It doesn't look like there is a schematic for the module, but there are schematics for the rest of the amp.
Cheers,
Zach |
|
|
| Kenshin |
| quote: | Originally posted by classd4sure
Hi Zach,
Look at the circuit! Total build cost (mass produced).. I'm not sure.. less than a dollar???
|
so cheap? what kind of gate driver did they use?
the schematic on the user's manual is too small and I can't recognize the words on the chip. |
|
|
| classd4sure |
| quote: | Originally posted by Kenshin
so cheap? what kind of gate driver did they use?
the schematic on the user's manual is too small and I can't recognize the words on the chip. |
Gate driver? In my opinion, they didn't have one. The user manual doesn't show the modulator at all, see this post:
http://www.diyaudio.com/forums/show...3572#post543572 |
|
|
| Kenshin |
Pat Allen:
my god, s64ami just used some 7414 paralleled to drive MOSFET gate! It's a fault design even for (robot race) motor drivers. And it's a part of a commercial audio D-amp design?!
perhaps BJT P+N emitter follower will be more powerful(but it needs design again)
At least,replace them using an 555,this can operate safely at 12V.
These are more than repairing. let's go to thread "Output devices" to talk. |
|
|
| charger_1 |
Hopfully Pat will be viewing this, as it pertains specifically about his reverse engineered S64AMI schematic. I found several flaws in it, errors in the circuit paths, and most notibly the parts values you listed are way off. If anyone has tried to breadboard this thing by that schematic, I would bet they didn't have much luck!
If I can figure out how to attach an image, I will put the corrected parts values and the schematic on here for everyone. I don't have that program that allows me to write PDF's, so Pat, If you wish to re-write this so it is in a better format, PLEASE DO!
What I did is take a screenshot of the PDF and altered it using a paint program so at least I had something. Sorry about the large JPG fellas! I didn't want to loose what little detail was left! |
|
|
| charger_1 |
Hopfully Pat will be viewing this, as it pertains specifically about his reverse engineered S64AMI schematic. I found several flaws in it, errors in the circuit paths, and most notibly the parts values you listed are way off. If anyone has tried to breadboard this thing by that schematic, I would bet they didn't have much luck!
If I can figure out how to attach an image, I will put the corrected parts values and the schematic on here for everyone. I don't have that program that allows me to write PDF's, so Pat, If you wish to re-write this so it is in a better format, PLEASE DO!
What I did is take a screenshot of the PDF and altered it using a paint program so at least I had something. Sorry about the large JPG fellas! I didn't want to loose what little detail was left!
I do have actual scans of both sides of the raw circuit board, so if anyone is interested in them, let me know and I will post them too.
Jeremy |
|
|
| classd4sure |
Hi Jeremy,
At the bottom when you're writing a post where it say's "attach file".... simple. Try saving your file as a PNG, no loss of quality, even in MSPaint, and very small size.
Can always zip it further if you have to.
Looking forward to seeing your corrections.
Take care. |
|
|
| pat allen |
| quote: | Originally posted by charger_1
Hopfully Pat will be viewing this, as it pertains specifically about his reverse engineered S64AMI schematic. I found several flaws in it, errors in the circuit paths, and most notibly the parts values you listed are way off. If anyone has tried to breadboard this thing by that schematic, I would bet they didn't have much luck!
If I can figure out how to attach an image, I will put the corrected parts values and the schematic on here for everyone. I don't have that program that allows me to write PDF's, so Pat, If you wish to re-write this so it is in a better format, PLEASE DO!
What I did is take a screenshot of the PDF and altered it using a paint program so at least I had something. Sorry about the large JPG fellas! I didn't want to loose what little detail was left! |
Hi i would like to see the original schematic for sure. BTW, i never said my reverse was perfect, and that i was unable to deliver correct parts values because they are unreadable. But at some point, i did make it work on my bread board, until some parts fryed because of a stupid mistake.
I didnt worked any further on it since i have done something quite better, and is under work for now.
Post anything you have about this module.
Thanks. |
|
|
| charger_1 |
OOPS! Not only did I upload the wrong schematic, I did it twice! (had troubles earlier)
Hopfully I got the right one this time.
The differences are that D8 and D9 are not zener diodes, and d7 is supposed to be 1n4937.
Sorry about the confusion! |
|
|
| charger_1 |
Pat,
I don't have the original schematic, but what I did is elaborate on yours. I was able to remove all the black epoxy and be able to retain all the parts, including their color bands and number codes (caps for instance). Talk about a tricky experiance! Then I was tracing through, using your schematic as refrence and noticed that the pathways weren't exactly correct (almost perfect, but I suppose since you used a mini torch to remove your black goo, things might have been burnt away) according to the circuit board traces. So I decided to expand on your start and make the necessary corrections. I also scanned the original board so if you would like to compare, to make sure I didn't miss anything, please feel free. I mapped out what goes where, it's orientation, etc. so nothing would get lost in the process of removing everything.
Hopfully the schematic will attach to one of these posts!
Jer |
|
|
| pat allen |
Hi, yes my board is a bit destroyed. Parts also. Keep us posted...
btw, you may send your stuff by e-mail and then i may be able to update and post...
patallen@aei.ca |
|
|
| pat allen |
Hi, C3= 4.7p and C7=100p ? i dont see why they are not the same value...just wondering.
And what about C11 ?? |
|
|
| charger_1 |
c3 and c7 are different probably because they run on different signal levels.... i don't know. C11 is part of a completely different circuit in the schematic. I wouldn't expect that to be the same just because it is attached to the same IC.
I pulled the parts off one at a time, documenting each one as it was removed. This is how the circuit was set up.
Friday I ran across this post, yesterday I removed and documented the components, and today I reverse engineered the schematic as it was laid out on the PCB. I haven't had any time as of yet to figure out why things are the way they are.
I'm goin gto attach the two board pictures with the components removed for those who would like them. Some of the solder mask tore off the board while removing the potting, however most of the detail is still clearly visible. |
|
|
| pat allen |
C3 and C7 are not on the same chip. They are in the signal path, one for the positive path and the other for the negative. If they are that much different, it may lead to a very unequal waveform, or it is justifed to make it equal. And bout C11, you just dont mention its value. It also seems that you dont have the same revison board as mine. Yours is 7 while mine is 6.
Do you found the burned component in yours? mine was both inverters that were shot. |
|
|
|
|
