Hello, this is my first "real" thread, outside of my introduction. I started working on some mods to my Adcom 545IIs, before locating this site, and its great resourses. My high level plans for these amps, can be found here.
After reading many threads here, I have started to have some concerns, about the Power Supply upgrades. Specificly, I read a post about how many commercial amps, are designed, such that the power supply sags, to help prevent the damage to output devices, by keeping them within SOA. The stock Adcom has a 500VA 80VCT transformer, with 56VDC rails. It utilizes three pairs of 2SA1492/2SC3856, per channel.(outputs are parrellel) Each channel has its own PS bridge, with 10K mfd per rail. (40Kmfd total)
The mod to the first amp, was to go with a 625VA 40-40 Avel transformer, using seperate bridges per rail, but would feed both channels. ie capacitance is shared for both channels. Plan was to use 60 or 80Kmfd. Amp 2 would then use the spare Adcom transformer to become a true dual mono, with 20Kmfd per rail.
Outside of the start-up surge, and current limiting that will probably be required, I am worried that I maybe putting the output transistors at risk. I do plan on using 4 ohm TMM and future 4 ohm surrounds. I have been reading posts and looking at the SOA graphs for these transistors, but not sure if I am in the safe realm. Also not sure, how to calculate the supply rail sag, that will happen at the same time. I looked at the schematic of the Rotel 991, which uses the same transistors, but 5 pairs, and has 73V rails.
I would appreciate, if someone could take a look at the datasheets, and see how much danger I'm in. It would be great if the calculation for both sag and soa could be shared as well. If any other data, like schematic or clarification on what I'm trying to say, let me know.
2SC3856 Datasheet
2SA1492 Datasheet
After reading many threads here, I have started to have some concerns, about the Power Supply upgrades. Specificly, I read a post about how many commercial amps, are designed, such that the power supply sags, to help prevent the damage to output devices, by keeping them within SOA. The stock Adcom has a 500VA 80VCT transformer, with 56VDC rails. It utilizes three pairs of 2SA1492/2SC3856, per channel.(outputs are parrellel) Each channel has its own PS bridge, with 10K mfd per rail. (40Kmfd total)
The mod to the first amp, was to go with a 625VA 40-40 Avel transformer, using seperate bridges per rail, but would feed both channels. ie capacitance is shared for both channels. Plan was to use 60 or 80Kmfd. Amp 2 would then use the spare Adcom transformer to become a true dual mono, with 20Kmfd per rail.
Outside of the start-up surge, and current limiting that will probably be required, I am worried that I maybe putting the output transistors at risk. I do plan on using 4 ohm TMM and future 4 ohm surrounds. I have been reading posts and looking at the SOA graphs for these transistors, but not sure if I am in the safe realm. Also not sure, how to calculate the supply rail sag, that will happen at the same time. I looked at the schematic of the Rotel 991, which uses the same transistors, but 5 pairs, and has 73V rails.
I would appreciate, if someone could take a look at the datasheets, and see how much danger I'm in. It would be great if the calculation for both sag and soa could be shared as well. If any other data, like schematic or clarification on what I'm trying to say, let me know.
2SC3856 Datasheet
2SA1492 Datasheet
SOA for audio is basically dissipation limited. You need to know the average supply voltage and output current and calculate transistor dissipation under worst case line conditions.
Voltage "sag" is a hard thing to predict. It depends on source impedance of the transformer and rectfiers and how much filtering there is. If all you've changed is the transformer VA rating, then the difference will only be whatever the difference in DCR of the windings.
Voltage "sag" is a hard thing to predict. It depends on source impedance of the transformer and rectfiers and how much filtering there is. If all you've changed is the transformer VA rating, then the difference will only be whatever the difference in DCR of the windings.
SOA for bipolar transistors is not just a dissipation matter
The dissipation rating of a bipolar device is only mantained up to a certain Vce, then thermal and secondary breakdown issues progressiveli reduce the dissipation [see SOA graphs in the datasheets]
+-56V DC are very conservative operating rails since they allow to drive fully reactive [90 deg lag on impedance] 4 ohm loads without exceeding DC SOA
In the other hand, for fully resistive loads +-80V DC allows to drive 4 ohm without exceeding DC SOA
Now, you may add some aditional cooling to the amplifier and push a bit these limits into AC SOA
For inductive loads [ie: real world loads] I don't recommend higher operating rails than +-72V since 100mS SOA would be exceeded
Of course, provided some sagging, no-load rails may be a bit higher than the values I mentioned
So, if the rest of the components in the circuit allow for it, you could use +-72V no load rails and let it sag to 56-66V with load and still mantain some realiability
Of course, you will be doing all the modifications at your own risk, I assume no liability and I can't guarantee any results [I've played to push devices outside their SOA and I have blown them sometimes]
The dissipation rating of a bipolar device is only mantained up to a certain Vce, then thermal and secondary breakdown issues progressiveli reduce the dissipation [see SOA graphs in the datasheets]
+-56V DC are very conservative operating rails since they allow to drive fully reactive [90 deg lag on impedance] 4 ohm loads without exceeding DC SOA
In the other hand, for fully resistive loads +-80V DC allows to drive 4 ohm without exceeding DC SOA
Now, you may add some aditional cooling to the amplifier and push a bit these limits into AC SOA
For inductive loads [ie: real world loads] I don't recommend higher operating rails than +-72V since 100mS SOA would be exceeded
Of course, provided some sagging, no-load rails may be a bit higher than the values I mentioned
So, if the rest of the components in the circuit allow for it, you could use +-72V no load rails and let it sag to 56-66V with load and still mantain some realiability
Of course, you will be doing all the modifications at your own risk, I assume no liability and I can't guarantee any results [I've played to push devices outside their SOA and I have blown them sometimes]
I'd recommend only going to 45VAC secondaries that will give you 63VDC after rectification/filtering. If you make the transformer and capacitor banks large enough your rails won't 'sag'. Sagging is a bad thing anyway, if you have tougher speaker loads (impedance like 2 ohms or less) it's going to sound a bit less dynamic. Going to somehow lesser rails would still keep the amp's stability in driving higly reactive loads.
You might also try soldering instead of using crimped connections for the power rails, it's a bit ugly and not too fiddling around friendly but you'll be surprised how much it matters when you hear the benefits. You have to use a 100W soldering iron or gun for the 'big' wires.
I don't know what the PSRR is in case of the Adcoms, you might go into some crosstalk issues above 10kHz if using one cap bank to feed both channnels, but if you're looking to biamp it won't matter for the one used for woofers.
Guess I should include these in my signature:
http://sound.westhost.com/power-supplies.htm
http://www.tnt-audio.com/clinica/ssps1_e.html
You might also try soldering instead of using crimped connections for the power rails, it's a bit ugly and not too fiddling around friendly but you'll be surprised how much it matters when you hear the benefits. You have to use a 100W soldering iron or gun for the 'big' wires.
I don't know what the PSRR is in case of the Adcoms, you might go into some crosstalk issues above 10kHz if using one cap bank to feed both channnels, but if you're looking to biamp it won't matter for the one used for woofers.
Guess I should include these in my signature:
http://sound.westhost.com/power-supplies.htm
http://www.tnt-audio.com/clinica/ssps1_e.html
Thanks for the replys.
Currently, and not changing now, because transformer has already been purchased
, I am staying with the 56V rails. I am only increasing the current capacity of the transformer, and increasing the the capacity of the storage bank. I was just concerned if there was sag designed into the Adcom, that the stronger power supply, would cause problems, by not sagging as much. Especially the 2 nd amp which will have a 500VA per channel. From your comments, I should by good with the 56V rails.
The TnT article, I have read many times. It is where I got the idea to improve on the PS. Also, this one provided allot of info. The ESP one, I have seen but have not read in detail. Will do so now.
The Adcom transformer, while having seperate wires for each channel, are actually connected together within the windings. I could wire the first amp like that, vs the bridges between rails. I would go to 80Kmfd if I did that. 40 a channel.
Any other comments or suggestions, are appreciated.
Currently, and not changing now, because transformer has already been purchased
, I am staying with the 56V rails. I am only increasing the current capacity of the transformer, and increasing the the capacity of the storage bank. I was just concerned if there was sag designed into the Adcom, that the stronger power supply, would cause problems, by not sagging as much. Especially the 2 nd amp which will have a 500VA per channel. From your comments, I should by good with the 56V rails.The TnT article, I have read many times. It is where I got the idea to improve on the PS. Also, this one provided allot of info. The ESP one, I have seen but have not read in detail. Will do so now.
The Adcom transformer, while having seperate wires for each channel, are actually connected together within the windings. I could wire the first amp like that, vs the bridges between rails. I would go to 80Kmfd if I did that. 40 a channel.
Any other comments or suggestions, are appreciated.
I've built regulated PSs for the Adcoms with really good results. Around 63v in and 56v out works with a 555II amp. You can also seperate the input PS from the output PS with two seperate transformers, a smaller one for the input(lower VA rating). That sounds really good too. The amp also has a protection circuit that limits the current to safe levels so dont worry about SOA issues.
Mark,
Some people on Audio Asylum, have indicated that going to a regulated PS for the input stage, would do more for the sound then the PS upgrade. I don't dispute this. The 545II chassis and standard high profile toroid, make it hard to just do that. With the 625VA Avel, I could fit another smaller toroid, on top of it. Avel doesn't make a 40-40 in the lower VA, so will need to look for others. I am also not sure where in the schematic to make the seperation. I will scan one channel of schematic and post on my website, latter this week. Maybe you can point out where I need to make the seperation. Does someone have a schematic or preferably a compact pcb for reg PS? Whith a reg ps on input, I do have a question, on a high level signal, where the output ps sags and the input doesn't, will that cause clipping?
Thanks again for the good input. Keep it coming!
Some people on Audio Asylum, have indicated that going to a regulated PS for the input stage, would do more for the sound then the PS upgrade. I don't dispute this. The 545II chassis and standard high profile toroid, make it hard to just do that. With the 625VA Avel, I could fit another smaller toroid, on top of it. Avel doesn't make a 40-40 in the lower VA, so will need to look for others. I am also not sure where in the schematic to make the seperation. I will scan one channel of schematic and post on my website, latter this week. Maybe you can point out where I need to make the seperation. Does someone have a schematic or preferably a compact pcb for reg PS? Whith a reg ps on input, I do have a question, on a high level signal, where the output ps sags and the input doesn't, will that cause clipping?
Thanks again for the good input. Keep it coming!
Hi, Johnsz -
I just performed a test with a 800va Avel toroid (as a first step in building up my 6 channel hybrid amp) 40-0-40 vac loaded at 150 watts and it only dropped from 61-0-61 vdc to 57-0-57 vdc under load. My AC mains voltage runs high - 120VAC or a bit more.
I was a little nervous because I am running a little close to the voltage limit of the 2 - 47,000 uF 63vdc capacitors I have & I will be running 2 stacked 800VA toroids in the eventual supply.
Turn on surge limiting was something I also tested. I put a 5 ohm at 25C NTC surge limiter across the terminals of a 25A AC relay with a delayed start up function. A little funky but it seemed to work pretty well at keeping turn on surges under control.
I just performed a test with a 800va Avel toroid (as a first step in building up my 6 channel hybrid amp) 40-0-40 vac loaded at 150 watts and it only dropped from 61-0-61 vdc to 57-0-57 vdc under load. My AC mains voltage runs high - 120VAC or a bit more.
I was a little nervous because I am running a little close to the voltage limit of the 2 - 47,000 uF 63vdc capacitors I have & I will be running 2 stacked 800VA toroids in the eventual supply.
Turn on surge limiting was something I also tested. I put a 5 ohm at 25C NTC surge limiter across the terminals of a 25A AC relay with a delayed start up function. A little funky but it seemed to work pretty well at keeping turn on surges under control.
thoriated,
Thanks for the input. If you don't mind, I have a couple of questions on your testing.
- How much of a load did you apply, when you got the 4V drop @ the output of the ps caps? The stock caps in the Adcom are 63V but my replacements are 80V. I was expecting 56V rails with 40V secondary.
- Did you try any tests without the surge protection? if so I'd like to know how bad of a surge issue you had.
- Can you share your surge protection circuit?
- Also curious in your amp design. Can you share any additional info?
Thanks.
Thanks for the input. If you don't mind, I have a couple of questions on your testing.
- How much of a load did you apply, when you got the 4V drop @ the output of the ps caps? The stock caps in the Adcom are 63V but my replacements are 80V. I was expecting 56V rails with 40V secondary.
- Did you try any tests without the surge protection? if so I'd like to know how bad of a surge issue you had.
- Can you share your surge protection circuit?
- Also curious in your amp design. Can you share any additional info?
Thanks.
I put about 100 ohms across the minus to the plus rail, so got a little less than 150 watts of dissipation. Also the four volt drop was with only one primary connected. With both connected this evening, I was getting about 3.1V drop from no load, so these toroids seem to have pretty low internal losses.
Without any startup limiting (this is with just one 800va transformer & 2 of the 47,000 uf caps - I plan to double both in the final version) , there was a very perceptible momentary dimming of an incandescent lamp I had on the same circuit and I could hear the transformer windings buzz for a second. I suspect that without limiting that the total supply I plan to use might be throwing circuit breakers.
With my initial prototype startup circuit, this effect was reduced by about two thirds & no transformer buzz, but by the time I finish tweaking it, I think I can make the startup surge almost imperceptible. I used a spare P&B opto coupled 25A 250VAC solid state relay paralleled with a Digi-Key KC005L Inrush Limiter (I think I can reduce the surge further by going with a higher resistance version of the surge limiter though). I create about a one second delay using a separate small transformer that has a 6.3V filament winding that I FW rectify then feed through a 150 ohm 1/2 watt resistor to a 6800uF 16V lytic across the logic level opto coupled inputs to the solid state relay. So, when I turn on, the inrush limiter does its thing, but it's deliberately underrated for the load to soften the surge, then after a second the solid state relay kicks in, essentially shunting the surge limiter which gets fairly warm but not hot like it usually would used as a standalone. This is a bit of a 'different' approach than I've seen before, but it's simple and seems to work ok.
The amp is planned to be a six channel hybrid HT amp, using a 6K11 compactron for the input and driver stage driving 3 pair of Toshiba 2SK1530 and 2SJ201 per channel. Besides prototyping the power supply, I'm doing the chassis now which will be a 3U rack mount style with substantial heatsink extrusions having 2.25" long fins with a 0.5" thick base plate good for maybe 0.6C/watt/channel. If I can get it to work out, I would like to set up four channels to be bridgeable in case I ever want to use it as a stereo biamp and maybe have a phantom center channel input for the two front channels. I think I have a scheme to control turn on and turn off pops and offsets which will use some new Normally Closed photomos SS relays that I hope I can get to shunt both halves of the output followers until the tubes are warmed up and to shut the output stage down quickly upon power down. Also, I plan to make each channel DC coupled but with a blocking cap at the input.
Without any startup limiting (this is with just one 800va transformer & 2 of the 47,000 uf caps - I plan to double both in the final version) , there was a very perceptible momentary dimming of an incandescent lamp I had on the same circuit and I could hear the transformer windings buzz for a second. I suspect that without limiting that the total supply I plan to use might be throwing circuit breakers.
With my initial prototype startup circuit, this effect was reduced by about two thirds & no transformer buzz, but by the time I finish tweaking it, I think I can make the startup surge almost imperceptible. I used a spare P&B opto coupled 25A 250VAC solid state relay paralleled with a Digi-Key KC005L Inrush Limiter (I think I can reduce the surge further by going with a higher resistance version of the surge limiter though). I create about a one second delay using a separate small transformer that has a 6.3V filament winding that I FW rectify then feed through a 150 ohm 1/2 watt resistor to a 6800uF 16V lytic across the logic level opto coupled inputs to the solid state relay. So, when I turn on, the inrush limiter does its thing, but it's deliberately underrated for the load to soften the surge, then after a second the solid state relay kicks in, essentially shunting the surge limiter which gets fairly warm but not hot like it usually would used as a standalone. This is a bit of a 'different' approach than I've seen before, but it's simple and seems to work ok.
The amp is planned to be a six channel hybrid HT amp, using a 6K11 compactron for the input and driver stage driving 3 pair of Toshiba 2SK1530 and 2SJ201 per channel. Besides prototyping the power supply, I'm doing the chassis now which will be a 3U rack mount style with substantial heatsink extrusions having 2.25" long fins with a 0.5" thick base plate good for maybe 0.6C/watt/channel. If I can get it to work out, I would like to set up four channels to be bridgeable in case I ever want to use it as a stereo biamp and maybe have a phantom center channel input for the two front channels. I think I have a scheme to control turn on and turn off pops and offsets which will use some new Normally Closed photomos SS relays that I hope I can get to shunt both halves of the output followers until the tubes are warmed up and to shut the output stage down quickly upon power down. Also, I plan to make each channel DC coupled but with a blocking cap at the input.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.