quasi said:
quasi said:
Quasi or Anyone that can explain,
I have been trying to figure out the voltage across the devices based on current flows, but the transistors in the current flow is confusing me. In "Example 2" I can follow R4 drop = 0.6 (.0013 X 470 = 0.611), but I am having a challenge figuring out how 1.2V was determined for the drop of T2 CE. I have looked at the datasheets and I have not been able with my limited knowledge to figure out how 1.2V was drop across T2. Is there a simple way to determine this referenced to Supply Rail +-?
As extension to this I wanted to apply this in all the other occurences, such as the the example you gave to Danko. I am hoping the answer to the "Example 2" question I posed just above I will be able to apply the same formula to figure this and all the other cases.
The goal I have is to put the circuit in a spreadsheet I will build to calculate currents, voltages, wattage when I change the Supply Rail voltage or plug in some alternate part value suggestions you have offered others. With this I can find out what part alternatives I may be able to use/test before settling on which I will use for a few amps I wish to build for me. It will also provide me a rough idea of currents/voltages expected should something go wrong whiel building or after fact when an amp has been working for while.
Regards,
John L. Males
Willowdale, Ontario
Canada
07 February 2006 01:05
Uh oh...sorry John...I mustta been typing asleep.
The correct calculation is
positive rail voltage - v(R4 & T2). with 70 volt rails this works out to be; 70 - 1.2 = 68.8v
1.2 volts is the total voltage from the 70 volt rail to the collector of T2. The voltage is set by the current through R12 and the base-emitter voltage of T6.
But hey what's 0.6 of a volt between amatuer DIY'ers ?
Cheers
The correct calculation is
positive rail voltage - v(R4 & T2). with 70 volt rails this works out to be; 70 - 1.2 = 68.8v
1.2 volts is the total voltage from the 70 volt rail to the collector of T2. The voltage is set by the current through R12 and the base-emitter voltage of T6.
But hey what's 0.6 of a volt between amatuer DIY'ers ?
Cheers
Why hasn't anyone considered the IRFP460 for the output stage of Quasi's amp? I think they should be suitable as well.
Just as an aside, I have used both IRFP460 and IRFP260 output devices in the erstwhile N-channel amps of Anthony Holton with very good results. While I used a small circuit from the upper rail to raise the gate voltage of the lower bank of output Mosfets which prevents cross-conduction, the IRFP260 with their larger capacitance sound great in the bass frequencies.
Just as an aside, I have used both IRFP460 and IRFP260 output devices in the erstwhile N-channel amps of Anthony Holton with very good results. While I used a small circuit from the upper rail to raise the gate voltage of the lower bank of output Mosfets which prevents cross-conduction, the IRFP260 with their larger capacitance sound great in the bass frequencies.
keypunch said:
My PCB-designing software puts thoose + marks there. I can "grab" the values of the components with them, and move the values away. When I print the picture, the + marks aren't there, but I just made a screenshot...
Mainly just for "put there a capacitor, I will decide the value later"
That "R" near the 1uF+100nF form an RC low-pass filter, to filter out the influence, coming from the high-current stage, the noise from the PSU (i plan to use an SMPS...). When I build the circuit, it's possible, that I solder there an electrolythic capacitor. I don't know yet
I can easily buy MPSA42 at the shop, where I used to buy electronic parts. I haven't studied the datasheet of MPSA42 and 2SC1845, but I hope, MPSA42 can replace this 2SCxxxx good. MPSA42 is a 300V TO92packaged device.
If I lower thoose resistors, I lower the I^2*R losses, but unfortunately I increase the thermal stability, as far as I know....
I choose to paralell 2x2W resistors, becouse the 5W ceramic resistors are wirewound, and the 2W types are metal-oxide type.
I don't really know, if the inductance of the ceramic resistors is a real problem in this circuit, or just I'm worrzing about it...
BTW, I'm dealing with switched mode powersupplies, where the inductance for such a resistor (for example in a current sensing position) is problem, maybe, this is why I choose 2W resistors.
I could buy for a good price 220uF/100V capacitors.
BTW, I have seen circuits, where a 100-200uF capacitor was soldered near EVERY MOSFET. It was also an audio amplifier.
I can also make it, if I see it's necessary.
I think, I made this, becouse 2uF/100V foil capacitors aren't at that shop, where I will buy the parts. Or I don't really remember
As I wrote a few lines upper, mostly I deal with SMPSs, so I think, using an 500V MOSFET in a 2x75V audio amplifier is too much waste
Danko, though the IRFP260 has some good Pd ability, but the T(off) time is really slow compared to T(on), about a 4/1 ratio that may lead to cross conduction issues in the upper audio frequency range. Most MOSFETs seem to have more of a 2/1 ratio of T(on)/T(off). The IRFP450 also has this 4/1 ratio
Hm, I haven't even think about this...
Thank you, for drawing my attention to this thing!
Well, (again
) as I told, I was dealing with switching mode, where the dead time solves this problem Thanks!
If everything goes well, I will get the parts on Friday
P.S.: This was a loooong post!
)
Hi guys,
Just a couple of thoughts from my end.
Regarding the input transistor here's the link to the 2SC1845 http://www.necel.com/discrete/images/pdf_e/TC-3236D.pdf. I think you'll find it's a pretty good transistor.
The audible difference between 0.47 and 0.33 or 0.22 resistors in the output stage is nil. The extra voltage drop at full power into 8 ohms amounts to just over 1 or 2 volts. So if it's a trade off between thermal stability, linearity and efficiency consider that in real terms 1 or volt is inaudible at full power. If you ran the amp into 2 ohms or 1 ohm (I had an email asking for this !!!) then source resistors could be a consideration. But the amp was never intended for this. Whilst it could be used for powerful PA / musical instrument or car boom boom applications, its birth into the world was intended to be as a quality hi-fi unit.
With respect to choice of FET's I have had many emails asking if this or that FET would work. The general answer is yes, and I doubt the cross-conduction would become a big issue when reproducing music. Yes there is some cross-conduction at 20khz but the output filter sucks a fair bit of current as well. I cannot comment about the sonic qualities of different FETs (I turn 47 this year). The reason I used IRFP450's is I have (or had) a bucket of them and I was able to buy some more at $2.20 US. If I did not have these then I would have chosen a FET with the best Ciss / power trade off.
So men (and women) build away, enjoy the FETs you have and put them to good use.
Cheers
PS; I also have a hundred or so IRF840's so stay tuned.
Just a couple of thoughts from my end.
Regarding the input transistor here's the link to the 2SC1845 http://www.necel.com/discrete/images/pdf_e/TC-3236D.pdf. I think you'll find it's a pretty good transistor.
The audible difference between 0.47 and 0.33 or 0.22 resistors in the output stage is nil. The extra voltage drop at full power into 8 ohms amounts to just over 1 or 2 volts. So if it's a trade off between thermal stability, linearity and efficiency consider that in real terms 1 or volt is inaudible at full power. If you ran the amp into 2 ohms or 1 ohm (I had an email asking for this !!!) then source resistors could be a consideration. But the amp was never intended for this. Whilst it could be used for powerful PA / musical instrument or car boom boom applications, its birth into the world was intended to be as a quality hi-fi unit.
With respect to choice of FET's I have had many emails asking if this or that FET would work. The general answer is yes, and I doubt the cross-conduction would become a big issue when reproducing music. Yes there is some cross-conduction at 20khz but the output filter sucks a fair bit of current as well. I cannot comment about the sonic qualities of different FETs (I turn 47 this year). The reason I used IRFP450's is I have (or had) a bucket of them and I was able to buy some more at $2.20 US. If I did not have these then I would have chosen a FET with the best Ciss / power trade off.
So men (and women) build away, enjoy the FETs you have and put them to good use.
Cheers
PS; I also have a hundred or so IRF840's so stay tuned.
Hi quasi,
I think part of ask about alternate transistors is in part due to cost, sourcing or others have "buckets" of X on hand (remind us of of anyone?) they want to use. I know at least for me I like to know the basics so I can research the net and datasheets to merge the info in my "grey matter" to make a decision. I suppose I think and behave "logical" as some birdy has suggested. This is why I asked about how to know what the voltage drop across a small/med level transistor CE voltage drop so I can do my own thing with the math and evaluation process.
I may of opened pandora's box regarding the cross conduction comment. Sorry if I did, it was not my intent. My research still seems unclear on cross conduction, so I mentioned in case someone else knew more about this. The context was respecting HiFi audio, not PA audio. I know with a couple of exceptions most MOSFETS, and for that matter BiPolars, seem to have specs that in theory imply cross conduction possibilities. See what some heavy research and all causes The point I was trying to make is maybe some devices are more prone than others perhaps overall design related and type of music aside.
I agree sonic qualities of various MOSFETs or other components is a highly debated topic. I know from the weeks of research I have done on various elements concerning amplifiers. By and large I have found much good information about others experience and thoughts on many of the likely asked questions you have had. For what it is worth, I personally think it is best to take the time to research and make one's own decision or ask questions based on the research or prior well understood experiences. Is this not what DIY is about?
One of those research efforts involved much time and very involved spreadsheet of specs and implied specifications of a large field of MOSFETs to consider for output devices. The list under consideration/evaluate was based on comments, specs, price and local availability. Interesting enough the IRF840 (and one of my favourites IRF740s is a prime candidate. So I will be very interested in your use and experiences with the IRF840. I really like to try out the IRF640N, but sadly I cannot source any here( ( ( The IRF540/IRF540N was also a very interesting consideration.
I know I asked about a board varient for TO-220 output device spacing. Then I withdrew that request. I did so after rethinking the thermal elements of TO-220 devices for an amp. I still hold by that opinion. It would be my suggestion with the IRF840's to still follow the TO-247 device to device spacing for the TO-220s that the IRF840 is a part of.
Regards,
John L. Males
Willowdale, Ontario
Canada
08 February 2006 03:13
I think part of ask about alternate transistors is in part due to cost, sourcing or others have "buckets" of X on hand (remind us of of anyone?) they want to use. I know at least for me I like to know the basics so I can research the net and datasheets to merge the info in my "grey matter" to make a decision. I suppose I think and behave "logical" as some birdy has suggested.
This is why I asked about how to know what the voltage drop across a small/med level transistor CE voltage drop so I can do my own thing with the math and evaluation process.I may of opened pandora's box regarding the cross conduction comment. Sorry if I did, it was not my intent. My research still seems unclear on cross conduction, so I mentioned in case someone else knew more about this. The context was respecting HiFi audio, not PA audio. I know with a couple of exceptions most MOSFETS, and for that matter BiPolars, seem to have specs that in theory imply cross conduction possibilities. See what some heavy research and all causes
The point I was trying to make is maybe some devices are more prone than others perhaps overall design related and type of music aside.I agree sonic qualities of various MOSFETs or other components is a highly debated topic. I know from the weeks of research I have done on various elements concerning amplifiers. By and large I have found much good information about others experience and thoughts on many of the likely asked questions you have had. For what it is worth, I personally think it is best to take the time to research and make one's own decision or ask questions based on the research or prior well understood experiences. Is this not what DIY is about?
One of those research efforts involved much time and very involved spreadsheet of specs and implied specifications of a large field of MOSFETs to consider for output devices. The list under consideration/evaluate was based on comments, specs, price and local availability. Interesting enough the IRF840 (and one of my favourites IRF740s is a prime candidate. So I will be very interested in your use and experiences with the IRF840. I really like to try out the IRF640N, but sadly I cannot source any here
( ( ( The IRF540/IRF540N was also a very interesting consideration.I know I asked about a board varient for TO-220 output device spacing. Then I withdrew that request. I did so after rethinking the thermal elements of TO-220 devices for an amp. I still hold by that opinion. It would be my suggestion with the IRF840's to still follow the TO-247 device to device spacing for the TO-220s that the IRF840 is a part of.
Regards,
John L. Males
Willowdale, Ontario
Canada
08 February 2006 03:13
hoangduongo said:
Hi hoangduongo,
To run off +/- 40v resistor R6 must be changed from 18K to 10K or can be left out and a link used instead.
The amp will not need as much gain so R17 should be replaced with a 22K resistor.
T1 & T5 can be replaced with BC550 or BC546 or similar, but watch for the different pin-outs. If you find a 2SCxxx in a TO92 with a Vce of 50v or more then that will do too.
All the MJE340's can be replaced with cheaper BD139. Likewise all the MJE350's can be replaced with BD140.
With 40 volt rails you will achieve about 55 watts into 8 ohms and about 90 into 4 ohms depending on your power supply.
Cheers & Good Luck
hoangduongo said:
hoangduongo,
Welcome.
I would suggest you take the time to read the entire thread. You will not only find the answers that address the one change suggested for a PSU of +-40V, you will find some important notes of corrections and input you may wish to consider for any other mods you have in mind that may be same or similar to ones you may be thinking about.
My last post was work in progress so I did not see it before I completed my last post. I think the IRF540 has a few endearing qualities as I indicated. I know of another n-channel project for IRF260N or IRF250N that a person elected to use IRF540N's. I beleive the person was very happy with the changes he made to a reduced PSU voltage and lower output driver count ( used 2 pairs rather than the 6 pairs suggested in design).
There is info in this thread that will enable you to determine how many pairs of devices you will need, based on output load, for a +-40V supply voltage and what current requirements you will need and verify if the transformer you wish to use is able to deliver to your requirements.
Regards,
John L. Males
Willowdale, Ontario
Canada
08 February 2006 03:45
08 February 2006 04:50 Some typo corrections
Oh well, the mighty quasi god has spoken whilst I created a reply!!!! lol lol
Quasi (bowing to the god),
Could in effect reduce power output using 44-0-44 by reducing the gain from about 150W to say 100W/8? Motivation is not the alternate transistor parts suggestion you made, just the ability to reduce the output power of the amp. I have access to some 44-0-44 transformers the until now was causing me to design for more power than I need for some amps I wanted to build. Umm how patience and not thinking of such a simple change could enable this design option.
Regards,
John L. Males
Willowdale, Ontario
Canada
08 February 2006 04:14
Quasi (bowing to the god),
Could in effect reduce power output using 44-0-44 by reducing the gain from about 150W to say 100W/8? Motivation is not the alternate transistor parts suggestion you made, just the ability to reduce the output power of the amp. I have access to some 44-0-44 transformers the until now was causing me to design for more power than I need for some amps I wanted to build. Umm how patience and not thinking of such a simple change could enable this design option.
Regards,
John L. Males
Willowdale, Ontario
Canada
08 February 2006 04:14
Hi John,
In theory the voltage across a transistor (Vce) can be as low as zero (or near enough) under saturation conditions with low current. I hope my second and more alert explanation makes more sense.
Cross conduction is an issue for Mosfets beacuse of the input capacitance, which without a discharge path can leave a FET turned on indefinately (in theory again). In this amp the only discharge path is via the pre-out stage and in particular R20 & R22. The discharge path can be improved by lowering the value of these resistors but then the current through this stage increases (4 volts / resistance) and with 150 volts rail to rail things can get very hot. So it's a compromise.
Transistors do not suffer as much because they are current driven, take the current away and they stop.
Cheers
In theory the voltage across a transistor (Vce) can be as low as zero (or near enough) under saturation conditions with low current. I hope my second and more alert explanation makes more sense.
Cross conduction is an issue for Mosfets beacuse of the input capacitance, which without a discharge path can leave a FET turned on indefinately (in theory again). In this amp the only discharge path is via the pre-out stage and in particular R20 & R22. The discharge path can be improved by lowering the value of these resistors but then the current through this stage increases (4 volts / resistance) and with 150 volts rail to rail things can get very hot. So it's a compromise.
Transistors do not suffer as much because they are current driven, take the current away and they stop.
Cheers
You may send holy gifts if you want but Quasi is just fine.
Seems we're both typing over the top of each other...reminds me of CB days....
The gain of the amp sets the input - output voltage ratio. It doesn't actually set the power capability of the amp....but I'm sure that's not what you meant.
If you can get the 44 volt transformers, to me that would offer the almost perfect amp setup. Use of IRF840/740/540 etc. with their lower Ciss and less heat through the second and third stages.....and still get near 150 watts into 8 ohms.....sounds pretty good to me (put a limiter on the volume control).
Cheers
Seems we're both typing over the top of each other...reminds me of CB days....
The gain of the amp sets the input - output voltage ratio. It doesn't actually set the power capability of the amp....but I'm sure that's not what you meant.
If you can get the 44 volt transformers, to me that would offer the almost perfect amp setup. Use of IRF840/740/540 etc. with their lower Ciss and less heat through the second and third stages.....and still get near 150 watts into 8 ohms.....sounds pretty good to me (put a limiter on the volume control).
Cheers
hoangduongo said:
hoangduongo,
Most of the credit of information goes to quasi. Much of whatI know has been from quasi or a few others hopping in to answer "a few" of my questions =)) As you can tell from my comments I have a few ideas of my own re MOSFETs parameters for output devices that I have found interesting. I am sure you will find when you read the thread those who have more knowledge than I about amps and this amp specifically.
I like to take a moment and thank all you have helped answered my inquiring mind questions. It has all been so very helpful and guiding. I know a few aspects I have not completley digested so more questions later in week or two, and I will try to break down in small digestable pieces.
I also need to take some time and update the project page I started to bring it up to date with the additional information since I first put it up.Regards,
John L. Males
Willowdale, Ontario
Canada
09 February 2006 10:02
09 February 2006 10:04 typo corrections
mailto:jlmales@softhome.net
Hi!
I hasuccesfully started the amplifier!
(After replacing the part, valued "39", which was meant to be 39pF, but I put there 39 ohm) )
I have one small problem: There's a typical class-B distorion at the zero-cross. And whatever I set the trim.pot at the BC546, nothing happens.
Can it be becouse I used a 100ohm pot. instean of the 200ohm ?
I hasuccesfully started the amplifier!
(After replacing the part, valued "39", which was meant to be 39pF, but I put there 39 ohm
) )I have one small problem: There's a typical class-B distorion at the zero-cross. And whatever I set the trim.pot at the BC546, nothing happens.
Can it be becouse I used a 100ohm pot. instean of the 200ohm ?
hello, quasi!
A tried to reverse the BC546, but nothing happened, whatever I turned the 200R trimpot. I also tried to change the 100R (I haven't got 200R at the moment) to 1k, but nothing happened.
By the way, the heatsink of the MJExxx transistors is cold. I think, that should be pretty warm, under normal conditions. So, I measured the voltage drop on R12 R15. On the R12 I measured the right voltage, about 0.6V, so that was OK. But on R15 I measured far more smaller voltage (I don't remember the exact value).
I'll do some more measurements, but I'm a little bit slow, I have to learn to my lessons....
Thanks, quasi!
A tried to reverse the BC546, but nothing happened, whatever I turned the 200R trimpot. I also tried to change the 100R (I haven't got 200R at the moment) to 1k, but nothing happened.
By the way, the heatsink of the MJExxx transistors is cold. I think, that should be pretty warm, under normal conditions. So, I measured the voltage drop on R12 R15. On the R12 I measured the right voltage, about 0.6V, so that was OK. But on R15 I measured far more smaller voltage (I don't remember the exact value).
I'll do some more measurements, but I'm a little bit slow, I have to learn to my lessons....
Thanks, quasi!
Hi,
First of all, what are your voltage rails?
If the heatsink of the of the MJExxx is cold it means you have no (or too little) current flowing through here.
You should check the components and for continuity for all of the second stage.
May I suggest you put 100 ohms 1 watt resistor or similar in line with the supply rails while you do your testing. This way if something goes wrong while your testing you will not blow up too many components.
Are you using my PCB layout or did you do your own?
Cheers
First of all, what are your voltage rails?
If the heatsink of the of the MJExxx is cold it means you have no (or too little) current flowing through here.
You should check the components and for continuity for all of the second stage.
May I suggest you put 100 ohms 1 watt resistor or similar in line with the supply rails while you do your testing. This way if something goes wrong while your testing you will not blow up too many components.
Are you using my PCB layout or did you do your own?
Cheers
You need to dimension your own power supply. You can use these rough guides;
Transformer rating = total power output of amplifier x 1.4. Example; 2 x modules @ 200 watts into 4 ohms will require a 560va transformer.
Use 5,000uF of capacitors per rail for every 100 watts
Example one module delivering 200 watts RMS into 4 ohms will need 2 x 10,000uF capacitors.
Always use a 25 amp diode bridge (they are too cheap to use anything else)
This layout is the one that I used for my amp. It may give you some ideas. http://www.diyaudio.com/forums/showthread.php?postid=819758#post819758
Use this schematic for your amp. It does not show the 4 extra FETs for the high power version but that does not matter. Just add the 4 to the PCB. This PCB is found here. http://www.diyaudio.com/forums/showthread.php?postid=797022#post797022
Cheers
Transformer rating = total power output of amplifier x 1.4. Example; 2 x modules @ 200 watts into 4 ohms will require a 560va transformer.
Use 5,000uF of capacitors per rail for every 100 watts
Example one module delivering 200 watts RMS into 4 ohms will need 2 x 10,000uF capacitors.
Always use a 25 amp diode bridge (they are too cheap to use anything else)
This layout is the one that I used for my amp. It may give you some ideas. http://www.diyaudio.com/forums/showthread.php?postid=819758#post819758
Use this schematic for your amp. It does not show the 4 extra FETs for the high power version but that does not matter. Just add the 4 to the PCB. This PCB is found here. http://www.diyaudio.com/forums/showthread.php?postid=797022#post797022
Cheers