A question for those who know high power design.
I have an old Yamaha M-2 power amp with a channel that died a couple of years ago. I "need" another amp to drive another JBL 2245H sub. I suppose I could try to repair the bad channel, but that wouldn't be keeping with the DIY spirit.
Not to mention that I haven't been able to find a legible schematic. I suppose I could just try replacing the 3 electrolytics on the boards.
The transformer is huge, although potted so I don't really know its true size. The amp was rated at 275 Watts per channel @ 8 ohms on 90 volt rails. It also has regulated 100 v rails for the front end. the heat sinks are not up to class A operation, but are a good size for a high power class AB amp. The (huge) main filter caps still seem OK, although at 21+ years old, I know I should replace them soon.
the output transistors are still good, 3 toshiba 2sa1051B/2sc2461B pairs per channel. Which leads me to my idea -
Would a modified Leach amp circuit work with 90V rails? I know about the double barrel version, but since that seems to have been designed around the unavailability of high voltage output devices, and I have the Low-TIM boards already...
I'd change R13/14, ala double barrel to accomodate the higher rails. Also substitute MJE15032 and MJE15033 for the 15030/15031 (Q16/17). I think I keep them in their SOA. Obviously, I'd need to use larger heat sinks on them and Q14/15.
Yamaha did not rate the amp for 4 ohm operation. Would this output stage survive bridged operation with an 8 ohm load? Home theater use with a fair amount of boost @ 20 Hz., not continuous PA type use. There is no room on the heat sinks for more T0-3 devices, but plenty of flat surface to substitute plastic ones.
Or would tuning it into a lowish bias Pass X amp clone make more sense? I sure hope that it is not some sort of collectors' item. It is the international version so I can change the rail voltage to 80 or 40-45 by changing the input voltage selector.
Thanks for your help, comments and suggestions.
I have an old Yamaha M-2 power amp with a channel that died a couple of years ago. I "need" another amp to drive another JBL 2245H sub. I suppose I could try to repair the bad channel, but that wouldn't be keeping with the DIY spirit.
Not to mention that I haven't been able to find a legible schematic. I suppose I could just try replacing the 3 electrolytics on the boards.The transformer is huge, although potted so I don't really know its true size. The amp was rated at 275 Watts per channel @ 8 ohms on 90 volt rails. It also has regulated 100 v rails for the front end. the heat sinks are not up to class A operation, but are a good size for a high power class AB amp. The (huge) main filter caps still seem OK, although at 21+ years old, I know I should replace them soon.
the output transistors are still good, 3 toshiba 2sa1051B/2sc2461B pairs per channel. Which leads me to my idea -
Would a modified Leach amp circuit work with 90V rails? I know about the double barrel version, but since that seems to have been designed around the unavailability of high voltage output devices, and I have the Low-TIM boards already...
I'd change R13/14, ala double barrel to accomodate the higher rails. Also substitute MJE15032 and MJE15033 for the 15030/15031 (Q16/17). I think I keep them in their SOA. Obviously, I'd need to use larger heat sinks on them and Q14/15.
Yamaha did not rate the amp for 4 ohm operation. Would this output stage survive bridged operation with an 8 ohm load? Home theater use with a fair amount of boost @ 20 Hz., not continuous PA type use. There is no room on the heat sinks for more T0-3 devices, but plenty of flat surface to substitute plastic ones.
Or would tuning it into a lowish bias Pass X amp clone make more sense? I sure hope that it is not some sort of collectors' item. It is the international version so I can change the rail voltage to 80 or 40-45 by changing the input voltage selector.
Thanks for your help, comments and suggestions.
I figured that only 3 pairs of output devices was pushing it, even for 8 ohms once I looked at the SOA of the recommended replacements. I guess I'll opt for 6-10 pairs of plastic package devices.
I'd like to keep the rail voltage up since my sub needs a ton of power. It seems to run out of steam occaisionally wnen driven by bridged Leach amps on 60 V rails with 78,000 uf in the supply. I will be moving the system to a much larger room soon, hence the need for another sub and amp.
I haven't had any power shortage driving the sub with a bridged Hafler DH-500. I suppose I coud clone that in this chassis, but those lateral MOSFETs aren't cheap.
I'd like to keep the rail voltage up since my sub needs a ton of power. It seems to run out of steam occaisionally wnen driven by bridged Leach amps on 60 V rails with 78,000 uf in the supply. I will be moving the system to a much larger room soon, hence the need for another sub and amp.
I haven't had any power shortage driving the sub with a bridged Hafler DH-500. I suppose I coud clone that in this chassis, but those lateral MOSFETs aren't cheap.
If the M2 is not rated for 4 ohm operation... then it is not rated for 8 ohm Bridge either.
If you are gung-ho on driving your Subs with high power, here's what I would do. Fix and sell the M2 on eBay. Buy an Adcom 555II from ebay and using a variac (or larger tranny) increase rails to about 85-87vdc. The front end will take it but the drivers will need to be replaced with MJE15034/5 or MJL1302 types.
I have done it many times. With Dual 800VA toroids, I get about 350 watts at clipping at 8 ohms (actually closer to 400 watts if I hook directly to AC mains socket) and 1100watts at 8 ohms bridged.
I also recommend an upgrade of the OP stage with MJ21193/4's. but stock should survive with 8 ohm loads.
K-
If you are gung-ho on driving your Subs with high power, here's what I would do. Fix and sell the M2 on eBay. Buy an Adcom 555II from ebay and using a variac (or larger tranny) increase rails to about 85-87vdc. The front end will take it but the drivers will need to be replaced with MJE15034/5 or MJL1302 types.
I have done it many times. With Dual 800VA toroids, I get about 350 watts at clipping at 8 ohms (actually closer to 400 watts if I hook directly to AC mains socket) and 1100watts at 8 ohms bridged.
I also recommend an upgrade of the OP stage with MJ21193/4's. but stock should survive with 8 ohm loads.
K-
Someone Posted:
"Would a modified Leach amp circuit work with 90V rails? I know about the double barrel version, but since that seems to have been designed around the unavailability of high voltage output devices, and I have the Low-TIM boards already..."
Probably not. Bipolar transistors have a destructive failure mechanism called second breakdown, which seems to be related to local hotspots becoming progressively hotter and conducting more current, a runaway heating condition. The applied voltage is critical, the hotspots do not seem to form below certain voltage/current levels. Devices are rated for second breakdown by stating the maximum current which should be allowed through the device with a given Vce on the device, and device specs also show curves for the second breakdown limitation. Exceeding these curves is usually fatal to the devices. This is why output devices are put in series or "cascode" mode, to divide the applied voltage and stay away from the second breakdown region of operation.
The load impedance and supply voltage determine what voltages and currents will exist, but in general I have found that, for a variety of real world speaker loads, exceeding about 60v rails for non-cascoded bipolar output stages is risky -- with even the best devices. I am a little conservative, and you can exceed this depending on your risk tolerance, but 90v rails are out of the question.
In the early 70s, Phase Linear had some 200wpc amps that did not use cascode connection, and tended to fail dramatically, taking speakers with them. Dynaco among others introduced it's ST4XX with cascode outputs, and these were fairly robust for the era. Threshold apparently even used cascoded stacks of 3 devices to achieve higher power. Leach's double barreled design is cascoded and probably the safest approach with bipolar output devices at higher power levels.
Mosfets output devices do not have this same second breakdown region and can be used with those voltage levels, provided other power dissipation limits are respected. Multiple devices would need to be paralleled, and possibly forced air cooling would be needed also.
"Would a modified Leach amp circuit work with 90V rails? I know about the double barrel version, but since that seems to have been designed around the unavailability of high voltage output devices, and I have the Low-TIM boards already..."
Probably not. Bipolar transistors have a destructive failure mechanism called second breakdown, which seems to be related to local hotspots becoming progressively hotter and conducting more current, a runaway heating condition. The applied voltage is critical, the hotspots do not seem to form below certain voltage/current levels. Devices are rated for second breakdown by stating the maximum current which should be allowed through the device with a given Vce on the device, and device specs also show curves for the second breakdown limitation. Exceeding these curves is usually fatal to the devices. This is why output devices are put in series or "cascode" mode, to divide the applied voltage and stay away from the second breakdown region of operation.
The load impedance and supply voltage determine what voltages and currents will exist, but in general I have found that, for a variety of real world speaker loads, exceeding about 60v rails for non-cascoded bipolar output stages is risky -- with even the best devices. I am a little conservative, and you can exceed this depending on your risk tolerance, but 90v rails are out of the question.
In the early 70s, Phase Linear had some 200wpc amps that did not use cascode connection, and tended to fail dramatically, taking speakers with them. Dynaco among others introduced it's ST4XX with cascode outputs, and these were fairly robust for the era. Threshold apparently even used cascoded stacks of 3 devices to achieve higher power. Leach's double barreled design is cascoded and probably the safest approach with bipolar output devices at higher power levels.
Mosfets output devices do not have this same second breakdown region and can be used with those voltage levels, provided other power dissipation limits are respected. Multiple devices would need to be paralleled, and possibly forced air cooling would be needed also.
"Would a modified Leach amp circuit work with 90V rails?"
Sure, just use MJ21195/96 outputs.
"I know about the double barrel version, but since that seems to have been designed around the unavailability of high voltage output devices, and I have the Low-TIM boards already..."
Don't know what you mean here, the double barrel version is designed for LOW VOLTAGE outputs. You couldn't use it anyway as the sink on the Yamaha isn't drilled for enough devices.
"Buy an Adcom 555II from ebay and using a variac (or larger tranny) increase rails to about 85-87vdc. "
Yeah, right. Which is higher? 6 X 250 or 8 X 150?
The Yamaha with 3 pairs of MJ21195/96 has 25% more safe area than an Adcom with 4 pairs of BD424/554, not to mention more heatsink area.
"Exceeding these curves is usually fatal to the devices"
Very true.
Consider a 90* reactive load, the absolute worst case. The power supply could provide 80V decaying to 60V.
The MJ21195/96 are 100% tested for 240W SOA at 80V, they can basically sink what the supply can source until the line fuse opens.
The lateral FETs used in a Hafler DH500 are only rated at 125W at 80V, an apt comparison as they run on similar rail voltages with the same number of output devices.
The Leach board has short circuit current limiting, I wouldn't worry about it.
Sure, just use MJ21195/96 outputs.
"I know about the double barrel version, but since that seems to have been designed around the unavailability of high voltage output devices, and I have the Low-TIM boards already..."
Don't know what you mean here, the double barrel version is designed for LOW VOLTAGE outputs. You couldn't use it anyway as the sink on the Yamaha isn't drilled for enough devices.
"Buy an Adcom 555II from ebay and using a variac (or larger tranny) increase rails to about 85-87vdc. "
Yeah, right. Which is higher? 6 X 250 or 8 X 150?
The Yamaha with 3 pairs of MJ21195/96 has 25% more safe area than an Adcom with 4 pairs of BD424/554, not to mention more heatsink area.
"Exceeding these curves is usually fatal to the devices"
Very true.
Consider a 90* reactive load, the absolute worst case. The power supply could provide 80V decaying to 60V.
The MJ21195/96 are 100% tested for 240W SOA at 80V, they can basically sink what the supply can source until the line fuse opens.
The lateral FETs used in a Hafler DH500 are only rated at 125W at 80V, an apt comparison as they run on similar rail voltages with the same number of output devices.
The Leach board has short circuit current limiting, I wouldn't worry about it.
Thanks, guys. I'll try the MJ21195/96 mod Leach. Too bad the usual suspects don't have them in stock. Arrow shows a 10 week lead time. Digikey has the similarly spec'd MJ21193/94 in stock tho.
I had a Phase Linear 400 in college. My brother had it when it blew. At least my speakers were OK.
Slowhands, I appreciate your caution. I'll set the transformer primary to give 80 volt rails. If I blow the outputs I'll go with MOSFETs the next time.
Digikey has the similarly spec'd MJ21193/94 in stock tho. I had a Phase Linear 400 in college. My brother had it when it blew. At least my speakers were OK.
Slowhands, I appreciate your caution. I'll set the transformer primary to give 80 volt rails. If I blow the outputs I'll go with MOSFETs the next time.
thanks Mark -
but i am curious. Why the difference? Most curves are nearly identical. About the only significant diffeence I see is the 80 V 1 second rating is .5 amp higher on the -95.
MJ21193 http://www.onsemi.com/pub/Collateral/MJ21193-D.PDF
High SOA: 2.5 A, 80 V, 1 Second
MAXIMUM RATINGS
VCEO 250 Vdc
VCBO 400 Vdc
VEBO 5 Vdc
IC 16 Peak30 Adc
IB 5 Adc
Total Power Dissipation PD 250 @ TC = 25°C
Derate Above 25°C 1.43 Watts W/°C
MJ21195 http://www.onsemi.com/pub/Collateral/MJ21195-D.PDF
High SOA: 3 A, 80 V, 1 Second
MAXIMUM RATINGS
VCEO 250 Vdc
VCBO 400 Vdc
VEBO 5 Vdc
IC 16 — Peak 30 Adc
IB 5 Adc
Total Power Dissipation PD 250 Watts@ TC = 25°C
Derate Above 25°C 1.43 W/°C
but i am curious. Why the difference? Most curves are nearly identical. About the only significant diffeence I see is the 80 V 1 second rating is .5 amp higher on the -95.
MJ21193 http://www.onsemi.com/pub/Collateral/MJ21193-D.PDF
High SOA: 2.5 A, 80 V, 1 Second
MAXIMUM RATINGS
VCEO 250 Vdc
VCBO 400 Vdc
VEBO 5 Vdc
IC 16 Peak30 Adc
IB 5 Adc
Total Power Dissipation PD 250 @ TC = 25°C
Derate Above 25°C 1.43 Watts W/°C
MJ21195 http://www.onsemi.com/pub/Collateral/MJ21195-D.PDF
High SOA: 3 A, 80 V, 1 Second
MAXIMUM RATINGS
VCEO 250 Vdc
VCBO 400 Vdc
VEBO 5 Vdc
IC 16 — Peak 30 Adc
IB 5 Adc
Total Power Dissipation PD 250 Watts@ TC = 25°C
Derate Above 25°C 1.43 W/°C
Its just my personal experience talking. I've used both many times and find the SOA of the 95/6 is different enough to make them usable at higher voltages. I love the 93/4 in my amps that dont have over 60 volt rails as they sound good and are very forgiving with short term high current outputs when used in multiples.
BobEllis said:Thanks, guys. I'll try the MJ21195/96 mod Leach. Too bad the usual suspects don't have them in stock. Arrow shows a 10 week lead time.
I had a Phase Linear 400 in college. My brother had it when it blew. At least my speakers were OK.
Slowhands, I appreciate your caution. I'll set the transformer primary to give 80 volt rails. If I blow the outputs I'll go with MOSFETs the next time.
How many pairs of 21195/96 do you want? I have many ....
BobEllis said:Thanks, guys. I'll try the MJ21195/96 mod Leach. Too bad the usual suspects don't have them in stock. Arrow shows a 10 week lead time.
I had a Phase Linear 400 in college. My brother had it when it blew. At least my speakers were OK.
Slowhands, I appreciate your caution. I'll set the transformer primary to give 80 volt rails. If I blow the outputs I'll go with MOSFETs the next time.
I have used the 21193/94/95/96 in the said Adcom 555II with no issues at +/- 100vdc rails!
You can be as conservative as you want which is sometimes good, yet from your earlier post, I sense the primary reason or interest you had was high power for a sub-woofer... in that scenario, you are fine with the 21195/96's even 3 pairs, heck I am willing to wager that the 3 pairs will be fine with a 4 ohm woofer no problems at 90v rails.
The 1 second SOA appears identical 3 amps at 80 volts for either device - perhaps they put in the wrong curve for the 93/94, or the 95/96 is more tolerant of short duration high current operation. At any rate, I'll trust your experience Mark. Thanks for the insight.
K-Amps, Please contact me off line robert.e.ellis (at) verizon.net, I'm interested in 6-10 pairs.
K-Amps, Please contact me off line robert.e.ellis (at) verizon.net, I'm interested in 6-10 pairs.
markp said:
Mark:
Your statement was about absolute voltage not SOAR. It read something like this, "The MJL21195/6 are higher voltage devices so you should stick with them."
Correction: They are not higher voltage devices... they are the same voltage devices in the absolute sense. They might handle higher currents for a given unit of time but that does not make them higher voltage devices.
Now please be a man and own up to the mistake you made instead of refuting it ad nauseum and sucking both of us in a peeing contest that I do not wish to be part of. I make many mistakes too, we are all human....
Now to your next statement: "They would not make two identical spec'ed transistors, would they?"
Define similarly spec'ed? Mototrola is one manufacturer I know for a fact makes large quantity parts and then grades them by yeilds. This grading is done per batch (pre-labelling) by non-destructive testing. They can test the secondary breakdown structures of the device and then categorize the batch as having higher SOAR or lower SOAR devices. And label them either as a 93/94 or a 95/95 part. Now they could have labelled everything as 21193/94 but then they would'nt be taking advantages of some parts that offered better SOAR parameters.
I did suggest to BobEllis (before) your post that I'd recommend him to upgrade the Adcom to 2119X devices. So to answer another one of your antagonistic remarks, 4X 250 is better than 3X 250. As far as heatsink area is concerned, it's importance is more of an issue with lower impedances which BobEllis did not mention that he needed.
Lastly the GFA-555II's heatsinks are much larger than the older 555's and can easily handle one woofer.... making heatsinking an issue when it is not is festering a non-productive argument. As suggested by slowhands, secondary breakdown is more of a primary concern than thermal equilibrium of the heatsinks.
Since I have been on the forum, I have learnt not to assume the other guy is an ignorant moron, it has saved me many a troublesome situations. I do not consider you one either and think you have given some good suggestions to our friend, I just feel you were a little more antagonistic than you needed to be.
regards,
K-
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.