| justonemoreamp |
Hi Paul, Aaron,
I used to work at a RF dealer, and because of this I could buy parts from them as a dealer of their products. This is back when the Punch HD models came out (~1992). It cost me max $6.00 for any of the cards used in that amp.
RF and PPI are the only Ceramic < Alumina> Based circuit card products I ever saw in after market car audio. I also saw ceramic card in the Weller TCP line of temp controlled solder stations., and a very small eq built into Fujitsu Ten/ Toyota decks < I still have this in a display on my dresser>
So the whole Ceramic Card thing was a engineering method to achieve a common set of goals.
First and foremost it was CHEAP. Secondly it was a easily handled product that simplified to overall design/build process. And third, it made product repair very quick and economical. But mostly it was CHEAP. In bulk quantities I am sure the costs were so low it was a no-brainer to have these made.
I also feel this "innovation" was just some new manufacturing methods to reduce costs just like MESHA is used currently by RF to reduce assembly costs. MESHA being nothing more then a adaptation of SMD technology. These ceramic card were early SMD technology and therefore promised CHEAP costs and manufacturing benefits.
Paul I think I can repair that damaged SIP card of yours. I stock all the semiconductors in large quantities, so its just a matter of getting the card off the board and performing the repairs. In fact I usually leave the cards installed to save time and just rework them in place.
I will answer the rest of your e-mail tonight for sure :)
Paul later my friend, Aaron nice to hear from you tonight:) |
|
|
| EnvisionAudio |
| quote: | Originally posted by justonemoreamp
Hi Paul, Aaron,
I used to work at a RF dealer, and because of this I could buy parts from them as a dealer of their products. This is back when the Punch HD models came out (~1992). It cost me max $6.00 for any of the cards used in that amp.
First and foremost it was CHEAP. Secondly it was a easily handled product that simplified to overall design/build process. And third, it made product repair very quick and economical. But mostly it was CHEAP. In bulk quantities I am sure the costs were so low it was a no-brainer to have these made.
I also feel this "innovation" was just some new manufacturing methods to reduce costs just like MESHA is used currently by RF to reduce assembly costs. MESHA being nothing more then a adaptation of SMD technology. These ceramic card were early SMD technology and therefore promised CHEAP costs and manufacturing benefits.
Paul later my friend, Aaron nice to hear from you tonight:) |
It is funny to think about these innovations this way. My current line of kits is all through hole so the average Joe can handle them without losing all the parts. However, my latest "innovation" is to go all SMD so that the kits are no longer "pieces parts". The key is to move towards assemblies. A lot of people just don't like to solder or assume it's some kind of "engineer's" thing. The simple fact is that SMD is far, far cheaper to assemble since the parts are placed 3-4 per second and all are soldered at the same time.
Rockford didn't go to SMD because it was better sounding, they did it because it was cheaper. But, we saw the transition from the ceramic "hybrid" designs towards full SMD because, frankly, it's an expensive setup to change over from through hole parts placers to chip shooters. So, like any good manufacturer, Rockford stuck with economies of scale and had the already-tooled ceramic boards made elsewhere.
For some reason, I thought Delco was making them. :confused: I throught I had seen them in dash clusters of GM products. The ones I was referring to had laser trimmed film resistors on board. |
|
|
| twon007 |
| quote: | Originally posted by Perry Babin
It looks like you have one transistor missing (bottom right-hand corner). If the board will accept solder, you can replace the missing transistor. If it was defective, the one directly above it has likely failed also.
If you're going to try to solder it, try to get some suggestions from justonemoreamp. He seems to have a lot of experiance with these boards. I've had some that would readily accept new solder and some that would not accept solder no matter what I tried (various iron temps, extra flux, silver bearing solder...).
If you're going to try to reproduce it, you could do so in a program like Cadsoft's Eagle. I was going to do it but I couldn't find a suitable connector to fit the holes in the main board so each connection would have to be made by a tiny piece of wire. That would be OK for a one-off piece but not for something the try to sell to others. The reproduction would probably be easier with through-hole components instead of SMD components.
Someone posted a full schematic for the board here but now I can't find it. Maybe someone else can remember enough about the post to locate it. |
yeah and it looks like a cap too. where is c23? |
|
|
| ppia600 |
| quote: | Originally posted by TO-3
Hey man- I looked at one of my PPIs this morning and found I gave out bad info. The gate resistors are supposed to be 47ohms(NOT 22) and the ground-tying resistor(directly on the outputs of the PWM chip Pins 11 & 14 is 2.2kohms) for the Z44s. Maybe it was the other MosFETs, though I don't recall thier P/N that was 22. Although, if your amp is working now, don't fix it. I apologize to all that may have used the 22s due to my info.
Whenever you replace one switcher, you need to replace them all. This is due to the amount of stress that the part inherits due to the higher amount of power it is subjected to.
Glad your amp is stable now. :) |
I have a 2150AM with blown power supply fets and all of my gate resistors read about 475 ohms each. That is .475 k ohms according to the meter. I see you talking about replacing them with 47 ohm resistors when swapping to the z44's, that seems kind of low for using on the gate of a mosfet, unless the switching IC is made to handle that load??? I just want to be sure of which gate resistors to use, since I plan on going with the plan of using the IRFZ44 mosfets.. |
|
|
| EnvisionAudio |
| quote: | Originally posted by ppia600
I have a 2150AM with blown power supply fets and all of my gate resistors read about 475 ohms each. That is .475 k ohms according to the meter. I see you talking about replacing them with 47 ohm resistors when swapping to the z44's, that seems kind of low for using on the gate of a mosfet, unless the switching IC is made to handle that load??? I just want to be sure of which gate resistors to use, since I plan on going with the plan of using the IRFZ44 mosfets.. |
No, 470 ohms is too high for a MosFet gate. There is no reason for these to be that high. MosFets draw nanomamperes on the gate pins - you're confusing them with regular bipolar transistors. You are most concerned with switching the FET on and off quickly. Even 47 ohms is high, but a good replacement, especially with the newer breed of MosFets. BTW, just move up to IRFZ48V and skip the Z44 - it has better peak current characteristics. |
|
|
| Perry Babin |
If I'm not mistaken, the 'AM' amplifiers used a transformer to drive the FETs. This would allow the use of a much higher value gate resistor. With a transformer, the gate drive of one bank will be at -3v before the others begin to turn on.
I'd try the new FETs with the original gate resistors but watch the temperature of the FETs. If they seem to be running hot, you may want to try reducing the value of the resistors. Unless someone with more experience can suggest a value, I'd go to 220 ohms (but only if there seems to be a problem). |
|
|
| EnvisionAudio |
| quote: | Originally posted by Perry Babin
If I'm not mistaken, the 'AM' amplifiers used a transformer to drive the FETs. This would allow the use of a much higher value gate resistor. With a transformer, the gate drive of one bank will be at -3v before the others begin to turn on.
I'd try the new FETs with the original gate resistors but watch the temperature of the FETs. If they seem to be running hot, you may want to try reducing the value of the resistors. Unless someone with more experience can suggest a value, I'd go to 220 ohms (but only if there seems to be a problem). |
Yes and no. While there is a transformer in the gate circuit, it is not a gate drive transformer. It is merely a common-mode filter (there is no secondary).
While your method seems OK for most amps, remember that the PPI amp bottom lid also serves as the transistor clamps. I don't recommend operating the amplifier without the lid attached unless you made a mechanical jig specifically for the purpose of troubleshooting of those amplifiers. |
|
|
| Perry Babin |
If that's the case and that's not a transformer, he definitely needs to go down to 100 ohms or less for the Z44s.
I assumed that he knew that he had to have the bottom cover in place (he's likely seen several of the other PPI repair threads). You can monitor the temperature through the heatsink. Generally, if the power supply end gets hot faster than the audio end, there's a serious problem. Some of the old Orions had a problem with the power supply heating up faster than the audio section.
I've had to make several sets of clamps for PPI amps (see attached). Some people think that using the bottom cover for a clamp is a great idea. I'm not one of them. |
|
|
| lumanauw |
| I've learned (the hard way) that 100ohm is much too big for IRFZ44. It will make cross conduction, making the whole SMPS mosfets blows up the whole set again and again. Better use 47ohm for IRFZ44's Rg. |
|
|
| Perry Babin |
100 ohm resistors are used with Z44s in a very large percentage of amplifiers. I'd probably use ~68 ohm resistors but 100 ohm should work.
Many of the Asian class D amplifiers use 100 ohm gate resistors with only 22 ohms for the dead time on an SG3525. If I'm not mistaken, PPI used 100 ohm dead time resistors in their older amplifiers which would provide even more dead time. |
|
|
| ppia600 |
Wow, I really opened a can of worms with that question. Yeah Perry, I might just use the z48v's like the ones I had to replace in the audiobahn.
First off, the original gate resistors on the ppi are 475 ohms. They are yellow(4), purple(7), brown(0), and the tolerance band is gold (5%). I'm pretty sure I'm not confusing bipolars with fets. I understand how sensitive most fets are at the gate, what I don't understand is how I can drop from a 470-475 ohm resistor to a 47 or even 22. That is quite a large jump. What in the architecture of a z44 for example makes the gate lead different than the original 25no5e's? |
|
|
| Perry Babin |
Brown is 1. :)
They were 470 ohms. 475 ohms would have had another band (yellow, violet, green, black, brown).
I tried to find the datasheet for the 25n05 but wasn't successful. I'd guess that a very low input capacitance and a lot of dead time allowed them to work with the 470 ohm resistors. I've only seen one other amp use values over 100 ohms (as far as I can remember). 10-100 ohms are the most common.
If you use Z48s, you need to use 47 ohm resistors (that's my suggestion anyway). |
|
|
| EnvisionAudio |
| quote: | Originally posted by ppia600
Wow, I really opened a can of worms with that question. Yeah Perry, I might just use the z48v's like the ones I had to replace in the audiobahn.
First off, the original gate resistors on the ppi are 475 ohms. They are yellow(4), purple(7), brown(0), and the tolerance band is gold (5%). I'm pretty sure I'm not confusing bipolars with fets. I understand how sensitive most fets are at the gate, what I don't understand is how I can drop from a 470-475 ohm resistor to a 47 or even 22. That is quite a large jump. What in the architecture of a z44 for example makes the gate lead different than the original 25no5e's? |
The 25N05E is a 50V 25Ampere device. A direct replacement for this obsolete device is an IRFZ34N. It's possible that PPI was intentionally trying to slow the turn-on of these FETs to reduce dv/dt - but I can't state that with any sort of proof. Who knows - maybe that is what the manufacturer of the FET recommended? Remember, this was at a time when using MosFets in the power supply was new. Maybe their research led them to believe that was a good thing.
Looking at the schematic...the dead time resistor is indeed 100 ohms. |
|
|
| ppia600 |
| quote: | Originally posted by EnvisionAudio
The 25N05E is a 50V 25Ampere device. A direct replacement for this obsolete device is an IRFZ34N. It's possible that PPI was intentionally trying to slow the turn-on of these FETs to reduce dv/dt - but I can't state that with any sort of proof. Who knows - maybe that is what the manufacturer of the FET recommended? Remember, this was at a time when using MosFets in the power supply was new. Maybe their research led them to believe that was a good thing.
Looking at the schematic...the dead time resistor is indeed 100 ohms. |
Yeah Perry, I was using the resistor color code wheel and it read 470, that is where I got the value from. The meter also read .470-.475k. (I wish it would just read 470 ohms but it doesn't have that range for some reason) The reason I said 470-475 ohms may be because my meter leads are crapping out, so sometimes it will fluctuate a little, but it could also be because of the 5% tolerance allowing the slight variance.
I still don't understand the reason for the metal can transistor guy saying the original gate resistors were 47 ohms..
Envision... where is the dead time resistor you're talking about located? And what is the exact reason they used the transformer/inductor between the two sets of gate resistors? Kind of interesting
I've also already ordered the z44 resistors, so which would be the optimum resistor to use at the gate now? Also, would it be a good idea to find some 1% tolerance resistors? Thanks for the help |
|
|
| EnvisionAudio |
| quote: | Originally posted by ppia600
Yeah Perry, I was using the resistor color code wheel and it read 470, that is where I got the value from. The meter also read .470 k. (I wish it would just read 470 ohms but it doesn't have that range for some reason) The reason I said 470-475 ohms may be because my meter leads are crapping out, so sometimes it will fluctuate a little, but it could also be because of the 5% tolerance allowing the slight variance.
I still don't understand the reason for the metal can transistor guy saying the original gate resistors were 47 ohms..
Envision... where is the dead time resistor you're talking about located? And what is the exact reason they used the transformer/inductor between the two sets of gate resistors? Kind of interesting
I've also already ordered the z44 resistors, so which would be the optimum resistor to use at the gate now? Thanks for the help |
The exact reason? I'd have to build their circuit to find out, but I think it's designed to reject common mode noise that might trigger both phases of the FET bank at the same time causing shoot-through. The schematic also shows a third winding connected to ground, bypassed with a 220uF cap to +12V to ostensibly reduce flux density in the core during dead time.
I have to admit - I've never seen this used before since the 3525 IC includes exclusive-OR logic lockouts to prevent double triggering of the outputs. Somebody must have thought it was a good idea, but it is obviously less of an issue that, say, transformer saturation or flux walking of a push-pull design; two mechanisms that definitely contribute to the cause of blown FETs.
The dead-time resistor is connected between pins 7 and 5 of the SG3525AN IC.
Regards,
Aaron Hammett |
|
|
| lumanauw |
Hi, Perry,
You're right, I forgot to include the dead time factor of the PWM IC. If the dead time is adjusted, 100ohm may be OK. My bad experience with 100ohm Rg on IRFZ44 is not with SG3525, but with TL494 with 0V at pin4.
Long time ago, I calculated the optimum Rg for IRFZ44 (with the formula found in one SMPS handbook, it turns to be about 12ohm. But practically, the totempole curent of the driver will be too big.
It is the discharge when OFF that makes the problem with too big Rg. IRFZ48 has bigger gate charge than IRFZ44, it will have different optimum Rg than IRFZ44. |
|
|
| ppia600 |
| Well I have the irfz44s on the way and some 100 ohm 1% tolerance resistors. Hopefully that is the right combo:smash: |
|
|
| ppia600 |
Okay, got all the z44's and gate resistors in. Double checked the outputs and found one bad 6490. Luckily I had a bag of them I ordered a couple of years ago and never used. Unfortunately I have no way to match them to the originals as far as hfe is concerned. Luckily, regular transistors are more forgiving than fets. I installed one of the new 6490's (central semi) and got it all regreased. So far, it works perfectly. I don't hear any unusual turn on or turn off noises either. The output section is getting warm at low power levels and the power supply fets aren't even making the sink warm to the touch. :D I'm currently running it at a 2 ohm mono load (yeah I know, not optimum) at a pretty low volume level to warm it up. If it can handle this stress test, it will surely do well at 2 ohms stereo. I'm glad the amps I love all have normal transistors for the output sections!
Thanks a lot for the help, guys! My friend is going to love this amp, he's never had a ppi and its going to make his four eclipse 6.5's shine like he's never heard. :) I hope he can truly appreciate the improved sound. |
|
|
| ppia600 |
| Update: Now that the amp has been running a little while I've swapped it to a 4 ohm load and the heatsink is staying about 115 farenheit. Looking good so far.:) |
|
|
| EnvisionAudio |
| Cool, mang - good to hear it! |
|
|
| HereticND |
I'm thinking this may be a long dead thread, but it's new to me :)
I'll be replacing the + fet rail, and one resistor on my 2150 m rev. D. So I should be getting the IRFZ44N fets, and the gate resistors should all be replaced too?, my R30 is burnt, that's the gate resistor right?
Also, do I need to test the two diodes that go into the fet circuits ie cr19 and cr20 from profuse's picture? How do I test them. One more thing, I have two other resistors that don't read at R5 and R38, are these on either side of the PWM? How do I test that and what does this tell you about why my amp shut down?
TIA
p.s. Part numbers would be helpfull as I don't remember the resistor band codes right now etc.
HereticND |
|
|
| Perry Babin |
What do you mean by "I'll be replacing the + fet rail"?
If the power supply isn't blown, you should leave it as it is.
R30 is an emitter resistor in the audio section. If it's defective, the transistor connected to it is also probably defective. If that's the case, you should replace all 4 emitter resistors and all 4 output transistors for that group of parallel components. |
|
|
| HereticND |
| quote: | | What do you mean by "I'll be replacing the + fet rail"? |
The mosfet on Q14 or Q15 isn't reading so I assumed from this thread that I had to replace all 15 on that side, which I think is the + side?
| quote: | | If the power supply isn't blown, you should leave it as it is. |
OK, if one mos fet is bad, then you have to replace them all right?
| quote: | | R30 is an emitter resistor in the audio section. If it's defective, the transistor connected to it is also probably defective. If that's the case, you should replace all 4 emitter resistors and all 4 output transistors for that group of parallel components. |
OK, if R30 is an emitter resistor, where are the gate resistors? Where are the transistors connected to it, cr 18, 19 & 20? Any guess on what to do with R5 and R38?
Thanks for the help! |
|
|
| Perry Babin |
Q14, Q15, Q16... are BJTs not MOSFETs. You need to replace at least 4 of them. BJTs have no gates. Instead, they have bases (base, collector, emitter instead of gate, drain source). Look up the datasheet for the transistors and it will show you the pin configuration.
It appears that Q14 is a driver that drives Q15-Q18). You can confirm this is you find that the emitter of Q14 goes to a group of resistors and from those resistors to the bases of the output transistors (Q15-Q18).
The emitter resistors are connected to the emitters of the output transistors.
If you have parallel components and one fails, you need to replace all of the components in the parallel group. Q15-Q18 are in parallel.
CR18-CR20 appear to be biasing diodes. It's unlikely that they have failed. You would need to set your meter to diode-check to test them in the board.
I can't find the resistors R5 and R38. |
|
|
| HereticND |
So what would be the part #'s for the Q14 - 18? Also the other resistors are to the right, towards the middle of the amp, on either side of a chip. Where's the code for resistor lines?
Thanks.
HereticND |
|
|
| Perry Babin |
What are the numbers on the transistors?
If you can't read the numbers, wipe some of the white heatsink onto the face of the transistor. When you wipe it off, you should be able to read the numbers clearly.
If the resistors are red, red, silver, gold, they are 0.22 ohm 5% tolerance. They are probably 2 watt resistors. When you order the replacements, measure the resistors in your amp and purchase resistors that are the same size.
If you want to learn more about the color code, read page #39 of my car audio site or search the web. There are many sites that have this information. |
|
|
|
