Re: The Table
John, it seems like you take a logical approach to calculating a safe output stage and from what you say I think your output stage decisions are sound.
As Richie00boy has suggested quite a bit of estimation goes on, but provided logical thinking is applied then things are fine without using all the formulas.
The differences between our "calculations" seem to be these in the main.
- I use the peak to peak voltage not the RMS voltage. In this way I work on the brief instantaneous power dissipated, so I use the 10 mSec SOAR.
- I use a reactive load model when I calculate the voltage across the FET and the current drawn by the load. As you can see with the sample chart, the voltage across the FET can actually exceed the rail voltage.
I do think though that I end up with an overly robust output stage, partcularly when I compare it to commercial units. If I seriously scrutinised the output stage in my Yamaha using my thinking, I doubt I would ever turn it on again.
Cheers
John, it seems like you take a logical approach to calculating a safe output stage and from what you say I think your output stage decisions are sound.
As Richie00boy has suggested quite a bit of estimation goes on, but provided logical thinking is applied then things are fine without using all the formulas.
The differences between our "calculations" seem to be these in the main.
- I use the peak to peak voltage not the RMS voltage. In this way I work on the brief instantaneous power dissipated, so I use the 10 mSec SOAR.
- I use a reactive load model when I calculate the voltage across the FET and the current drawn by the load. As you can see with the sample chart, the voltage across the FET can actually exceed the rail voltage.
I do think though that I end up with an overly robust output stage, partcularly when I compare it to commercial units. If I seriously scrutinised the output stage in my Yamaha using my thinking, I doubt I would ever turn it on again.
Cheers
Re: The Table
Hi Quasi,
Thanks for your reply. Yes I tend to be locical, but as we all know logic without the correct facts is not very logical. Ergo my review of how I calculated SOAR for me. I thank you for reviewing my approach to SOAR.
As I noted I try to keep even the peaks within the output device RMS ratings. I do this in the calculations via a P-P as well to ensure this. I take this approach as I do not hve the knowledge or skills to evaluate the reactive components a load will present to the output devices. Maybe one day I will understand the reactive elements of a load so I can add that to the suite of formulas I use to check the output drive design and PSU design. I have heard as RichiBoy has stated there are often elements that one cannot calculate nor measure all that easy and I understand that reality. I, like you have done, strive to apply as many known calculations to validate a design to be within the (safe) range to avoid surprises.
Interesting you mentioned that sometimes the voltage across the MOSFET will exceed the PSU rail voltage at times. I will have to look at your sample chart to see this as I did not notice. It may provide me some insight if I need to change the safety margin the calculations in my spreadsheet use. Easy spreadsheet change I just change on cell and all the MOSFET calculations are redone taking the change into account as I am sure you do with your spreadsheets for values you like to adjust to evaluate design changes.
BTW, I decided long ago to use a Tc of 70C as one of my safety margins as well. Part of this rational was also to take into account peaks and reactive elements I had no way of calculating. I found generally speaking for the other parameters values I used there was not much difference to 50C or 60C for my purposes when I reviewed these operating temperatures. I know such a choice is usually for amps pushed very hard or used for long hours in DJ or like suitations. My amps will be used for classical, jazz, pop, soft rock, alternative, et al music which means I should have the safety margin via Tc as one other elements to those elements I do not have the skills to calculate, or are elements one cannot really calculate.
Thanks again Quasi. I do have one or two other simple questions, but they can wait until the New Year. Have a great Holiday Season with the family and friends. I suppose now it is later evening Christmas Eve in your part of the globe. All The Best.
Regards,
John L. Males
23 December 2005 08:06
Willowdale, Ontario
Canada
Hi Quasi,
Thanks for your reply. Yes I tend to be locical, but as we all know logic without the correct facts is not very logical. Ergo my review of how I calculated SOAR for me. I thank you for reviewing my approach to SOAR.
As I noted I try to keep even the peaks within the output device RMS ratings. I do this in the calculations via a P-P as well to ensure this. I take this approach as I do not hve the knowledge or skills to evaluate the reactive components a load will present to the output devices. Maybe one day I will understand the reactive elements of a load so I can add that to the suite of formulas I use to check the output drive design and PSU design. I have heard as RichiBoy has stated there are often elements that one cannot calculate nor measure all that easy and I understand that reality. I, like you have done, strive to apply as many known calculations to validate a design to be within the (safe) range to avoid surprises.
Interesting you mentioned that sometimes the voltage across the MOSFET will exceed the PSU rail voltage at times. I will have to look at your sample chart to see this as I did not notice. It may provide me some insight if I need to change the safety margin the calculations in my spreadsheet use. Easy spreadsheet change I just change on cell and all the MOSFET calculations are redone taking the change into account as I am sure you do with your spreadsheets for values you like to adjust to evaluate design changes.
BTW, I decided long ago to use a Tc of 70C as one of my safety margins as well. Part of this rational was also to take into account peaks and reactive elements I had no way of calculating. I found generally speaking for the other parameters values I used there was not much difference to 50C or 60C for my purposes when I reviewed these operating temperatures. I know such a choice is usually for amps pushed very hard or used for long hours in DJ or like suitations. My amps will be used for classical, jazz, pop, soft rock, alternative, et al music which means I should have the safety margin via Tc as one other elements to those elements I do not have the skills to calculate, or are elements one cannot really calculate.
Thanks again Quasi. I do have one or two other simple questions, but they can wait until the New Year. Have a great Holiday Season with the family and friends. I suppose now it is later evening Christmas Eve in your part of the globe. All The Best.
Regards,
John L. Males
23 December 2005 08:06
Willowdale, Ontario
Canada
Hi quasi,
That's been my chief concern with Yamaha amplifiers for years. I used to fix them under warranty. Their pro amps are well made curiously. When a Yamaha home product blew, it went from outputs to diff pair pretty well. They didn't used to silkscreen the component side which made them a favorite to fix under warranty.
Not to throw darts at Yamaha only, most consumer gear is rather weakly constructed. That may be because my "normal" was the 70's Marantz product.
-Chris
That's been my chief concern with Yamaha amplifiers for years. I used to fix them under warranty. Their pro amps are well made curiously. When a Yamaha home product blew, it went from outputs to diff pair pretty well. They didn't used to silkscreen the component side which made them a favorite to fix under warranty.
Not to throw darts at Yamaha only, most consumer gear is rather weakly constructed. That may be because my "normal" was the 70's Marantz product.
-Chris
Re: Hey well done!
Hello quasi and keypunch, I'll be very happy to tell you about my amp.
The power supply is made with a E-I trf that is rated at 450va (acording to my maths), primary is 120v and secondary is 2x44v.
I used 2 bridge rectifiers, one for each rail and filters are 10,000uF nippon/chemicon + 6,800uF nichicon, each rated at 63vdc.
The amp itself uses 2 IRFP450 per rail, I left the space open so I could put another if needed. Right now I'm using it to drive 1 12inch 8ohms per channel, and it sounds very good.
I'm looking forward to build 2 15inch 8ohm next year for the amp, and leave the 12 inches on the old AKAI. I'll let you know about it
Happy holydays (and listening) to everyone, Merry Christmas and Happy New Year.
quasi said:
Hello quasi and keypunch, I'll be very happy to tell you about my amp.
The power supply is made with a E-I trf that is rated at 450va (acording to my maths), primary is 120v and secondary is 2x44v.
I used 2 bridge rectifiers, one for each rail and filters are 10,000uF nippon/chemicon + 6,800uF nichicon, each rated at 63vdc.
The amp itself uses 2 IRFP450 per rail, I left the space open so I could put another if needed. Right now I'm using it to drive 1 12inch 8ohms per channel, and it sounds very good.
I'm looking forward to build 2 15inch 8ohm next year for the amp, and leave the 12 inches on the old AKAI. I'll let you know about it
Happy holydays (and listening) to everyone, Merry Christmas and Happy New Year.
Re: Hey well done!
LuigiDJ,
Your amplifier configuration is very similar to what I am working out some details on. I have a few toriods that are 44-0-44, but measure 46-0-46 due to the AC voltage here being 125VAC. I have two types of these, a 926VA and 1127VA. Originally I was going to run 4 of quasi amp modules per one of these toroids. I have been doing various calculaltions upon calculations and have decided that I wanted more margin for reactive and low ohms load condition for peaks. I decided 4 modules even with the same PSU approach you have (of one rectifier on each side of the rail as well I am planning on each amp module having its own rectifier set and filter capacitor set) that 4 modules, even 3, was pushing matters far too thin.
I also have a good number of PC mount Nippon/Chemicon 10,000UF 63V capacitors. For the amps I make using the 44-0-44V toroid I plan on using 3x10,000uF on each side of the rail. When I can find 30-0-30 toroids I will use 2x10,00uF each side of the rail.
Using a 44-0-44 transformer secondary it seems a bit tight to be using capacitors rated at 63V. The rail voltage could be about 62V a side of rail before MOSFET losses. I am still working on sourcing capacitors that are at least 75V, but ideally 100V to improve lifespan of the capacitors when using the 44-0-44 toroids I have.
You may want to check your rail voltages under no load, to confirm if the rail voltage of your PSU exceeds the 63V rating of your capacitors.
You may also want to consider placing a third pair of IRFP450's in to reduce the risk of blowing one of the two pairs of IRFP450's. I recall basically one should plan that each pair of IRFP450s can handle 65W at 8 ohms, and 130W at 4 ohms.
You have a double design issue. The first and most basic is with a 44-0-44 transformer you will likely have about a 35V RMS rail assuming 3 pairs of IRFP450's. This will mean based on the rail voltage the possible power output, even if just a peak, is about 160W. Your two pairs will mean you are short by about 30W based on the 44-0-44 transformer you are using. Secondly, complicating this is you only have two pairs of output devices which translates to a rail loss less with two pairs of output devices, as opposed to three. This means the effective rail voltage will be higher, therefore higher power level the PSU will be able to deliver. This will translate into about 175W. So in essence you will be short about 45W of power handling of your pair of output drivers with a 44-0-44 transformer. This will likely result in a higher Tc running temperature for your output devices, IRFP450's, will run at or extra heatsinking/cooling is provided to disipate the extra heat from the two pair of IRFP450's.
For a 450VA transformer you can easily calculate the current demand and limits for given speaker impedance load. Please bear in mind that with stereo, both channels will likely peak about same time at same level. The math behind the current demand is fairly straight forward ohms law, but ignores the reactive elements of a speaker load. If your 450VA transformer calculation is total VA for both secondaries you will need to divide 450VA by two to know the VA rating for each side of the centre taped secondary can handle. This would mean 450/2/44=5.11Amps a per side of rail. At about 36V rail voltage per side of rail 36V/5.11=7.05 ohms is the minimum impedance a speaker can have for this transformer. It is likely a 8 Ohm rated speaker will dip below 8 ohms often and long periods of time, in fact well below 7 ohms. This will possibly strain the pair of IRFP450's you have and also heating up the transformer in the process causing other problems.
This above assumes a few things. First, if you push the amps a bit, common for woofer drivers to demand or need more power. Second, your music has frequent peaks beyond 130W for an 8Ohm load. The speaker will likley have an impedance over most of the frequency range it will operate more in the 6-4 ohm range, not the rated speaker impedance of 8 ohms. Speaker impedance is reactive, not resistive, which will place greater demands on the amplifier than the resistive 8 ohms rating of the voice coil often viewed as only element of of a speaker load.
Initially the amps I will make will have to cope with the passive crossover speaker networks, therefore making the load more of a (reactive) challenge to the amp. This is part of reason I decided to limit the number of modules to two per toroid. Later I will move to an active crossover. I suspect I will move to active in stages. This means splitting off the woofer actively from the mid range and tweeter still on own passive crossover. The next phase will be to make all crossovers active. I might try a passive crossover between pre-amp and amplifiers just as an initial experiment. The amps I will build with the 44-0-44 toroids will serve as the woofer amps on the imitial active crossover cut over.
Regards,
John L. Males
24 December 2005 03:14/03:18
Willowdale, Ontario
Canada
LuigiDJ said:
LuigiDJ,
Your amplifier configuration is very similar to what I am working out some details on. I have a few toriods that are 44-0-44, but measure 46-0-46 due to the AC voltage here being 125VAC. I have two types of these, a 926VA and 1127VA. Originally I was going to run 4 of quasi amp modules per one of these toroids. I have been doing various calculaltions upon calculations and have decided that I wanted more margin for reactive and low ohms load condition for peaks. I decided 4 modules even with the same PSU approach you have (of one rectifier on each side of the rail as well I am planning on each amp module having its own rectifier set and filter capacitor set) that 4 modules, even 3, was pushing matters far too thin.
I also have a good number of PC mount Nippon/Chemicon 10,000UF 63V capacitors. For the amps I make using the 44-0-44V toroid I plan on using 3x10,000uF on each side of the rail. When I can find 30-0-30 toroids I will use 2x10,00uF each side of the rail.
Using a 44-0-44 transformer secondary it seems a bit tight to be using capacitors rated at 63V. The rail voltage could be about 62V a side of rail before MOSFET losses. I am still working on sourcing capacitors that are at least 75V, but ideally 100V to improve lifespan of the capacitors when using the 44-0-44 toroids I have.
You may want to check your rail voltages under no load, to confirm if the rail voltage of your PSU exceeds the 63V rating of your capacitors.
You may also want to consider placing a third pair of IRFP450's in to reduce the risk of blowing one of the two pairs of IRFP450's. I recall basically one should plan that each pair of IRFP450s can handle 65W at 8 ohms, and 130W at 4 ohms.
You have a double design issue. The first and most basic is with a 44-0-44 transformer you will likely have about a 35V RMS rail assuming 3 pairs of IRFP450's. This will mean based on the rail voltage the possible power output, even if just a peak, is about 160W. Your two pairs will mean you are short by about 30W based on the 44-0-44 transformer you are using. Secondly, complicating this is you only have two pairs of output devices which translates to a rail loss less with two pairs of output devices, as opposed to three. This means the effective rail voltage will be higher, therefore higher power level the PSU will be able to deliver. This will translate into about 175W. So in essence you will be short about 45W of power handling of your pair of output drivers with a 44-0-44 transformer. This will likely result in a higher Tc running temperature for your output devices, IRFP450's, will run at or extra heatsinking/cooling is provided to disipate the extra heat from the two pair of IRFP450's.
For a 450VA transformer you can easily calculate the current demand and limits for given speaker impedance load. Please bear in mind that with stereo, both channels will likely peak about same time at same level. The math behind the current demand is fairly straight forward ohms law, but ignores the reactive elements of a speaker load. If your 450VA transformer calculation is total VA for both secondaries you will need to divide 450VA by two to know the VA rating for each side of the centre taped secondary can handle. This would mean 450/2/44=5.11Amps a per side of rail. At about 36V rail voltage per side of rail 36V/5.11=7.05 ohms is the minimum impedance a speaker can have for this transformer. It is likely a 8 Ohm rated speaker will dip below 8 ohms often and long periods of time, in fact well below 7 ohms. This will possibly strain the pair of IRFP450's you have and also heating up the transformer in the process causing other problems.
This above assumes a few things. First, if you push the amps a bit, common for woofer drivers to demand or need more power. Second, your music has frequent peaks beyond 130W for an 8Ohm load. The speaker will likley have an impedance over most of the frequency range it will operate more in the 6-4 ohm range, not the rated speaker impedance of 8 ohms. Speaker impedance is reactive, not resistive, which will place greater demands on the amplifier than the resistive 8 ohms rating of the voice coil often viewed as only element of of a speaker load.
Initially the amps I will make will have to cope with the passive crossover speaker networks, therefore making the load more of a (reactive) challenge to the amp. This is part of reason I decided to limit the number of modules to two per toroid. Later I will move to an active crossover. I suspect I will move to active in stages. This means splitting off the woofer actively from the mid range and tweeter still on own passive crossover. The next phase will be to make all crossovers active. I might try a passive crossover between pre-amp and amplifiers just as an initial experiment. The amps I will build with the 44-0-44 toroids will serve as the woofer amps on the imitial active crossover cut over.
Regards,
John L. Males
24 December 2005 03:14/03:18
Willowdale, Ontario
Canada
I,am back, thanks, people !
Quasi is 100% right. I finished power amp, just 2ch. for probe. I,get 42V @ 1Khz on 8 Ohm, 37V@ 1Khz on 4 Ohm. Toroid is 600VA , 2x 56 AC, but power comsuption is 2A@220V AC. Calculate this , is 440W from Mains. This means that toroid is only for 1 CH8 or 4 Ohm. It is my opinion. FAQ. What is the voltage for 10 Fets, & ofcourse what is a idle current , for 6Fets & 10Fets?
Quasi is 100% right. I finished power amp, just 2ch. for probe. I,get 42V @ 1Khz on 8 Ohm, 37V@ 1Khz on 4 Ohm. Toroid is 600VA , 2x 56 AC, but power comsuption is 2A@220V AC. Calculate this , is 440W from Mains. This means that toroid is only for 1 CH8 or 4 Ohm. It is my opinion. FAQ. What is the voltage for 10 Fets, & ofcourse what is a idle current , for 6Fets & 10Fets?
Hi Zeonrider
Good to see your amp is working.
The idle current should be set for 30mA per FET pair. So for 6 FETs it will be 90mA and for 10 FETs it will be 150mA.
Provided your power supply is well built the 600va transformer will power 2 modules easily. This because under music conditions the average load on the transformer is very low.
I have some questions, I hope you don't mind;
Did you measure the transformer voltage (2 x 56v) while it was loaded or unloaded (no power or full power)
What are your DC rails with zero signal from the amp?
What size capacitors are you using?
Is the voltage you measured RMS?
Cheers
Good to see your amp is working.
The idle current should be set for 30mA per FET pair. So for 6 FETs it will be 90mA and for 10 FETs it will be 150mA.
Provided your power supply is well built the 600va transformer will power 2 modules easily. This because under music conditions the average load on the transformer is very low.
I have some questions, I hope you don't mind;
Did you measure the transformer voltage (2 x 56v) while it was loaded or unloaded (no power or full power)
What are your DC rails with zero signal from the amp?
What size capacitors are you using?
Is the voltage you measured RMS?
Cheers
OK
The voltage is 2x56AC no load.Yes RMS . 2x70DC per rail, but I rised voltage on variac when I mesaured big drop. 2x33000uF TOWA. YOU didn't write me voltage for 10 Fets (2xAC or 2xDC).I,ll change capacitors 2x25000uF SPRAGUE, because I think the TOWA has big iner resistance.Please answerd me as soon as posible .
The voltage is 2x56AC no load.Yes RMS . 2x70DC per rail, but I rised voltage on variac when I mesaured big drop. 2x33000uF TOWA. YOU didn't write me voltage for 10 Fets (2xAC or 2xDC).I,ll change capacitors 2x25000uF SPRAGUE, because I think the TOWA has big iner resistance.Please answerd me as soon as posible .
Thanks Quasi !
Just litle more questions. What is the power (VA) of power transformer 1 or 2 CH for 10 Fets, and what happend with adj. BIAS .I couldn't regulate it , and allways had crossover. Yes something happened when I changed Tr8 BC 546 to BD 139, but only when I turned trim-pot in full one side.
Just litle more questions. What is the power (VA) of power transformer 1 or 2 CH for 10 Fets, and what happend with adj. BIAS .I couldn't regulate it , and allways had crossover. Yes something happened when I changed Tr8 BC 546 to BD 139, but only when I turned trim-pot in full one side.
Hi Zeonrider,
To run 85 volts with 10 FETs you will need at least a 1000va transformer to run a stereo setup.
Regarding setting the bias current, please make sure that the transistor pin outs are connected correctly. They are shown reversed for a BC546 on the PCB because I intended this transistor to mount under the PCB.
You said something happened. Does the amp still work?
Cheers
To run 85 volts with 10 FETs you will need at least a 1000va transformer to run a stereo setup.
Regarding setting the bias current, please make sure that the transistor pin outs are connected correctly. They are shown reversed for a BC546 on the PCB because I intended this transistor to mount under the PCB.
You said something happened. Does the amp still work?
Cheers
T8 Pinouts
Quasi,
I am assuming the T8 pinouts are labled on the Layout PCB are correct, i.e. as long as one knows the pinout of what one is using for T8 and pays attention to the layout markings on the PCB one will be fine?
Regards,
John L. Males
27 December 2005 22:12
Willowdale, Ontario
Canada
quasi said:
Quasi,
I am assuming the T8 pinouts are labled on the Layout PCB are correct, i.e. as long as one knows the pinout of what one is using for T8 and pays attention to the layout markings on the PCB one will be fine?
Regards,
John L. Males
27 December 2005 22:12
Willowdale, Ontario
Canada