With all due respect to David Hafler, I agree with a number of Mr. Nelson's remarks.
The fact is that speed distortions are more dependent on Group Delay and its behavior far beyond the audio band, up to 1 MHz and beyond. It is not enough to reduce the time Propagation Delay to almost zero in the audio range, it is important to keep this level further. But with this in the XL-280 amplifier there are big problems. See Attached Files.
Attachments
For those who do not understand basic things in electronics, please ignore my posts and do not flood like
Musings on amp design... a thread split.
Musings on amp design... a thread split.
After reading woody, HaflerDH500Fan, and Bob Cordell’s comments on the thermal resistance of the DH-200 heatsink, I thought I would add some empirical data.
I have mounted a MOSFET amplifier to the heatsinks from a scrapped Hafler DH-200 amplifier. The output stage of these amplifiers consist of a pair of IRFP240 and IRFP9240 FETs with 0.22 ohm source resistors. The FETs were biased at about 100 mA each. The rail voltage was ±68 volts. The amplifier was operated with shorted input, and a 4-ohm load. I mounted an Omega surface mount thermistor feeding an Omega meter to measure heatsink temperature.
The heatsink temperature was 17.2ºC prior to power-up. The heatsink temperature equilibrated out at 34.0ºC after about 60 minutes. I re-measured the bias for each FET, they varied between 97.7 and 107.7mA. The dissipation for the four FETs varied between 6.55 – 7.22W. The dissipation summed to 27.43W. if I did the math right, this will calculate out to a thermal resistance of 0.61ºC/W.
I suspect it may make a difference as to were on the heatsink the measurement is taken, but I am unaware of a standard, so I just used one of the available holes, see attached photo.
Cheers,
ceulrich
I have mounted a MOSFET amplifier to the heatsinks from a scrapped Hafler DH-200 amplifier. The output stage of these amplifiers consist of a pair of IRFP240 and IRFP9240 FETs with 0.22 ohm source resistors. The FETs were biased at about 100 mA each. The rail voltage was ±68 volts. The amplifier was operated with shorted input, and a 4-ohm load. I mounted an Omega surface mount thermistor feeding an Omega meter to measure heatsink temperature.
The heatsink temperature was 17.2ºC prior to power-up. The heatsink temperature equilibrated out at 34.0ºC after about 60 minutes. I re-measured the bias for each FET, they varied between 97.7 and 107.7mA. The dissipation for the four FETs varied between 6.55 – 7.22W. The dissipation summed to 27.43W. if I did the math right, this will calculate out to a thermal resistance of 0.61ºC/W.
I suspect it may make a difference as to were on the heatsink the measurement is taken, but I am unaware of a standard, so I just used one of the available holes, see attached photo.
Cheers,
ceulrich
Attachments
If you want the real thing, with latest mods, at this price a lot less than just the chassis and power transformer cost…
Hafler DH-220 Power Mosfet Amplifier - Very nice condition For Sale - Canuck Audio Mart
Hafler DH-220 Power Mosfet Amplifier - Very nice condition For Sale - Canuck Audio Mart
Interesting measurements. The 0.61 °C/W thermal resistance would be about the minimum required to support the Hafler DH220 listed power output of 115 watts into 8 ohm based on the assumptions listed in my previous note: (60°C-25°C)/(0.5*115 watts) = 0.61 °C/W.
BC547 / BC556 in PC19 card
I am curious if anyone has used (sorted / matched / installed) BC546/BC556 in PC-19 driver cards? I gripe and moan about the limits of the stock'ish PC-19... even recapped, with far ~ s u p e r i o r ~ better sounding parts. I will admit I have not completely given up... lingering curiosity is driving me.
I have a set of ~35 year old PC19 boards [from my original DH-220, pulled from storage] I built with these matched transistors, which sounded / tested exceptionally well... wanting to build and test a few extra driver PC boards... *while rebuilding this amp and these driver cards + plus a few; more while I am at it.
*Image shown, from late 80's build - early 90's usage... this amp was "kind of rough", early in my amp building days... Yes, 10 gauge Monster Cable in the B+... this amp had big balls... During the "amp clinics" of the day, embarrassed many a big name amp...
Notice ["large" PS storage] 330uf in the driver stage and 100uf on the outputs.
Some things never change = remain the same... ~ are required ~
I am curious if anyone has used (sorted / matched / installed) BC546/BC556 in PC-19 driver cards? I gripe and moan about the limits of the stock'ish PC-19... even recapped, with far ~ s u p e r i o r ~ better sounding parts. I will admit I have not completely given up... lingering curiosity is driving me.
I have a set of ~35 year old PC19 boards [from my original DH-220, pulled from storage] I built with these matched transistors, which sounded / tested exceptionally well... wanting to build and test a few extra driver PC boards... *while rebuilding this amp and these driver cards + plus a few; more while I am at it.
*Image shown, from late 80's build - early 90's usage... this amp was "kind of rough", early in my amp building days... Yes, 10 gauge Monster Cable in the B+... this amp had big balls... During the "amp clinics" of the day, embarrassed many a big name amp...
Notice ["large" PS storage] 330uf in the driver stage and 100uf on the outputs.
Some things never change = remain the same... ~ are required ~
Attachments
Last edited:
I've always matched these transistors in these amps. Makes a massive difference. I designed a BJT matching jig that does a pair at a time. It is the only way to match pairs as they must be kept at the exact same temperature.
Your replacement caps on the PCB are too large. The wire is ridiculous.
-Chris
Your replacement caps on the PCB are too large. The wire is ridiculous.
-Chris
an opinion, with possible amplifier envy??
~ Chris we often stand on entirely different sides of nearly every equation ~
Sadly, I have come to expect this ...kind of sounds personal... like a person [or people?] spending too much time on the work bench, fixing other peoples problems and not enough time enjoying music through even a decent HiFi set-up. I was under the impression you did not even own a(ny) Hafler amp(s).
~ Give me a break ~
**I built this amp as shown, 35-40 years ago. IT ..WAS RIDICULOUS.., FOR THE TIME (compared to) sitting next to my buddies mighty Levinson ML-3 driving my Maggies... I also used it on my Quad ESL-63, sitting next to Futterman OTL3 amps... RIDICULOUSLY GOOD SOUNDING!!!
This amp (as shown) was my daily driver, even next to numerous VENERABLE vacuum tube amps... of the day.
~ Chris we often stand on entirely different sides of nearly every equation ~
Sadly, I have come to expect this ...kind of sounds personal... like a person [or people?] spending too much time on the work bench, fixing other peoples problems and not enough time enjoying music through even a decent HiFi set-up. I was under the impression you did not even own a(ny) Hafler amp(s).
~ Give me a break ~
**I built this amp as shown, 35-40 years ago. IT ..WAS RIDICULOUS.., FOR THE TIME (compared to) sitting next to my buddies mighty Levinson ML-3 driving my Maggies... I also used it on my Quad ESL-63, sitting next to Futterman OTL3 amps... RIDICULOUSLY GOOD SOUNDING!!!
This amp (as shown) was my daily driver, even next to numerous VENERABLE vacuum tube amps... of the day.
Hi John,
Just an observation, that's all.
I had to live through the silly wire phase as well. Monster was king. I saw all kinds of destroyed gear where folks ripped out factory wire and caps, installing replacements too large to fit.
Honestly, smaller gage wire would perform exactly as good - and it fit the terminals! Your losses were not great with factory skinny wire, and slightly larger would eliminate all slight problems. But no, massive wire was the rule and the fad.
Many designers used their heads and used wire and capacitors that were the correct size. They were largely ignored because the magazine articles focused on the extreme. This was before the internet and before bulletin boards.
Upgrade artists were about the worst of the lot, being 100% fad driven without any logic to guide them. We have the same issues today of course.
Do I have an issue with this? Yup, I sure do. But I have no personal issues with you at all. What I do have issues with is destroyed equipment or stuff that performs at a substandard level after the great expensive work is done.
I rebuilt a Hafler DH101 with Teflon PCBs and dual power supplies. It was unstable, and suffered from devices running too hot. The huge dual power supplies were nuts and unnecessary. I redesigned it with one of the existing supplies and at the end it was reliable with better performance. The customer liked it a lot more too. To be honest, the original DH101 was a better unit. I have rebuilt those as well but was able to retain the original design while improving performance.
My definition of better performance is both measured and customer reports. Better performance is achieved by applying engineering rules and common sense. It costs a lot less to do it that way too.
Besides, the Hafler equipment was designed as a cost effective product that delivered good performance at a good price. Intelligent application of techniques can improve performance a lot - economically. What people had done with this equipment was over the top and ineffective. Spending the same money as the upgrades cost including labour, the customer was better off just buying a better unit.
Right now the best upgrade for some Hafler amplifiers is the excellent work done by Bob Cordell and Rick here. Want to improve something? Apply engineering and common sense. Throwing expensive parts and bigger wire at it didn't help performance very much at all. I know because I have measured every amp I came into contact with for decades. I always measure performance "as found", during improvements, and the final product.
I'm just very tired of fixing messed up stuff. Done it for years. Misinformation is my target, not you.
-Chris
Just an observation, that's all.
I had to live through the silly wire phase as well. Monster was king. I saw all kinds of destroyed gear where folks ripped out factory wire and caps, installing replacements too large to fit.
Honestly, smaller gage wire would perform exactly as good - and it fit the terminals! Your losses were not great with factory skinny wire, and slightly larger would eliminate all slight problems. But no, massive wire was the rule and the fad.
Many designers used their heads and used wire and capacitors that were the correct size. They were largely ignored because the magazine articles focused on the extreme. This was before the internet and before bulletin boards.
Upgrade artists were about the worst of the lot, being 100% fad driven without any logic to guide them. We have the same issues today of course.
Do I have an issue with this? Yup, I sure do. But I have no personal issues with you at all. What I do have issues with is destroyed equipment or stuff that performs at a substandard level after the great expensive work is done.
I rebuilt a Hafler DH101 with Teflon PCBs and dual power supplies. It was unstable, and suffered from devices running too hot. The huge dual power supplies were nuts and unnecessary. I redesigned it with one of the existing supplies and at the end it was reliable with better performance. The customer liked it a lot more too. To be honest, the original DH101 was a better unit. I have rebuilt those as well but was able to retain the original design while improving performance.
My definition of better performance is both measured and customer reports. Better performance is achieved by applying engineering rules and common sense. It costs a lot less to do it that way too.
Besides, the Hafler equipment was designed as a cost effective product that delivered good performance at a good price. Intelligent application of techniques can improve performance a lot - economically. What people had done with this equipment was over the top and ineffective. Spending the same money as the upgrades cost including labour, the customer was better off just buying a better unit.
Right now the best upgrade for some Hafler amplifiers is the excellent work done by Bob Cordell and Rick here. Want to improve something? Apply engineering and common sense. Throwing expensive parts and bigger wire at it didn't help performance very much at all. I know because I have measured every amp I came into contact with for decades. I always measure performance "as found", during improvements, and the final product.
I'm just very tired of fixing messed up stuff. Done it for years. Misinformation is my target, not you.
-Chris
I agree. Great post Chris.
While it is intuitive to many, fat power supply wires are neither necessary nor highly effective in improving performance. In fact, in the right setting, they may not be as good.
I do believe in larger reservoir capacitors to the extant that they can fit. These days, you can get capacitors with 3X the capacitance of the old computerlytics used in the Hafler in the same space. 30,000 uf per rail is a good number. This is often realized on a small board that will fit in the space occupied by the old capacitors. Often it will consist of 3 X 10,000 uF on each rail. Net ESR and ESL tend to be quite a bit lower with such an arrangement. Some such power supply boards, including nice rectifiers, can be found on eBay for on the order of $50.
I believe just as importantly that one should have 1000 uF or so rail caps on the output board right at the power transistors. This is the approach taken in the DH-220C.
Finally, with the above approach, smaller AWG power supply wire can actually make things better, since its modestly small resistance allows the reservoir and board capacitors to act as a bit of a pi network.
Cheers,
Bob
While it is intuitive to many, fat power supply wires are neither necessary nor highly effective in improving performance. In fact, in the right setting, they may not be as good.
I do believe in larger reservoir capacitors to the extant that they can fit. These days, you can get capacitors with 3X the capacitance of the old computerlytics used in the Hafler in the same space. 30,000 uf per rail is a good number. This is often realized on a small board that will fit in the space occupied by the old capacitors. Often it will consist of 3 X 10,000 uF on each rail. Net ESR and ESL tend to be quite a bit lower with such an arrangement. Some such power supply boards, including nice rectifiers, can be found on eBay for on the order of $50.
I believe just as importantly that one should have 1000 uF or so rail caps on the output board right at the power transistors. This is the approach taken in the DH-220C.
Finally, with the above approach, smaller AWG power supply wire can actually make things better, since its modestly small resistance allows the reservoir and board capacitors to act as a bit of a pi network.
Cheers,
Bob
in the wrong hands (things) happen
unfounded, inaccurate, with unbelievable bias...
~ That's BIG JOHN ; to you Chris!!! ~
To think that you can drop a super broad bandwidth circuit into a rats nest like a common DH101 with untidy wiring mess, potential shield, and grounding issues... is an overwhelming oversight. 101 a better unit are you even serious Chris?
My MC1 Teflon, as shown was in an ideal enclosure, with (only one switch, in the signal path) uniform, tidy twisted wiring... only phono and line level feeds, no tape loops, no external processing loops, switches... UNKNOWNS you clearly encountered and mentioned. I used mine alongside and ARC SP-10, with Futterman OTL (wide-bandwidth amps) at time prone to parasitic oscillation... driving Mark Levinson ML-3 (Maggies) and ML-2 (on Quads 57s & 63s) Never had a single problem with mine with a broad range of revealing / demanding gear...
THE MC1 IN THE RIGHT HANDS WAS ON PAR WITH THE BEST OF THE DAY! *the notion to think you could re-design it, I am sorry is simply ridiculous, borders blatant self promotion, reeks of misunderstanding?? or misinformation and who knows what level of gear you associated with it (30 years ago?).
Think about it, you tell others *not to touch* many stock designed circuits and yet your assertions imply otherwise, in promoting your vision and skills, to challenge and surpass industry leaders, frankly for which you principally seem to serve in repairing their gear when called upon and needed. Just saying' a little hypocritical, yet you have relied on this stream of mal-adjusted and broken gear for your own lively-hood.
~ Even a little humility, with a larger dose of clarity might be helpful ~
No point in arguing this further, I bring it up hopefully to steer past (ongoing labeling, bordering petty) grossly misguided misinformation, hopefully to help others here, unfamiliar with LEGACY product(s) that a few have only touched, many have never seen or heard. More importantly some (even you?) clearly do not understand.
Sorry to jerk the thread sideways here. A mod? started the topic,
unfounded, inaccurate, with unbelievable bias...
~ That's BIG JOHN ; to you Chris!!! ~
To think that you can drop a super broad bandwidth circuit into a rats nest like a common DH101 with untidy wiring mess, potential shield, and grounding issues... is an overwhelming oversight. 101 a better unit are you even serious Chris?
My MC1 Teflon, as shown was in an ideal enclosure, with (only one switch, in the signal path) uniform, tidy twisted wiring... only phono and line level feeds, no tape loops, no external processing loops, switches... UNKNOWNS you clearly encountered and mentioned. I used mine alongside and ARC SP-10, with Futterman OTL (wide-bandwidth amps) at time prone to parasitic oscillation... driving Mark Levinson ML-3 (Maggies) and ML-2 (on Quads 57s & 63s) Never had a single problem with mine with a broad range of revealing / demanding gear...
THE MC1 IN THE RIGHT HANDS WAS ON PAR WITH THE BEST OF THE DAY! *the notion to think you could re-design it, I am sorry is simply ridiculous, borders blatant self promotion, reeks of misunderstanding?? or misinformation and who knows what level of gear you associated with it (30 years ago?).
Think about it, you tell others *not to touch* many stock designed circuits and yet your assertions imply otherwise, in promoting your vision and skills, to challenge and surpass industry leaders, frankly for which you principally seem to serve in repairing their gear when called upon and needed. Just saying' a little hypocritical, yet you have relied on this stream of mal-adjusted and broken gear for your own lively-hood.
~ Even a little humility, with a larger dose of clarity might be helpful ~
No point in arguing this further, I bring it up hopefully to steer past (ongoing labeling, bordering petty) grossly misguided misinformation, hopefully to help others here, unfamiliar with LEGACY product(s) that a few have only touched, many have never seen or heard. More importantly some (even you?) clearly do not understand.
Attachments
guys, guys, guys (the era)
That WAS nearly 40 years ago!!!!!! Did you ever experiment??
Even a little... THAT is how we learn. How I learn.
~ at least it was a clean build ~ in my early days.
*why do people struggle to be more positive?
AS PER, I follow, have intuitively and by modeling
applied common sense / Bob's thinking in many ways,
it is simply good engineering...
It just makes good sounding gear. Consistently so.
If it makes y'all feel better... A picture
(of a starting point, PS build)
of a stock'ish rebuild for a buddy,
which I made 2 years ago.
That WAS nearly 40 years ago!!!!!! Did you ever experiment??
Even a little... THAT is how we learn. How I learn.
~ at least it was a clean build ~ in my early days.
*why do people struggle to be more positive?
AS PER, I follow, have intuitively and by modeling
applied common sense / Bob's thinking in many ways,
it is simply good engineering...
It just makes good sounding gear. Consistently so.
If it makes y'all feel better... A picture
(of a starting point, PS build)
of a stock'ish rebuild for a buddy,
which I made 2 years ago.
Attachments
Hi Chris,
Why teflon? At the time I was designing pcb's at HP, Cyanate Ester Resin (HT-2) was the premium dielectric. HP had two pcb fabs, PRCO in Loveland, CO and Boeblingen, DE. HT-2 was co-developed with Nelco. That was in the mid 90's. We never needed to use HT-2 in our designs but they were not RF or high power designs. I have used FR-4 into the GHz ranges on 10Gbit/s SerDes links. Like I said makes no sense to me
iirc teflon had it's inherent problems with de-lam ??P.S. There is nothing wrong with the DH200C I took to Dave's, something was a miss with his setup to causes the DH200C and his heavily modified EB unit to smoke a resistor, something up with the grounding.
Happy Holidays
Rick
Last edited:
Hi Chris,
Why teflon? At the time I was designing pcb's at HP, Cyanate Ester Resin (HT-2) was the premium dielectric. HP had two pcb fabs, PRCO in Loveland, CO and Boeblingen, DE. HT-2 was co-developed with Nelco. That was in the mid 90's. We never needed to use HT-2 in our designs but they were not RF or high power designs. I have used FR-4 into the GHz ranges on 10Gbit/s SerDes links. Like I said makes no sense to me
P.S. There is nothing wrong with the DH200C I took to Dave's, something was a miss with his setup to causes the DH200C and his heavily modified EB unit to smoke a resistor, something up with the grounding.
Happy Holidays
Rick
The last I knew, Panasonic Megtron 6 was the top PCB dielectric for high-frequency designs. I used it in 40 Gb/s and 100 Gb/s designs for undersea multi-wavelength fiber optic systems. A key factor that influenced some high-frequency dielectrics was the move to ROHS compliance, which requires significantly higher reflow temperatures. That doomed some of the Rogers dielectrics that we previously used.
Cheers,
Bob
Hi Guys,
I wasn't clear I guess.
The DH-101 I worked on was modified by John in this case, and he used a Teflon PCB. It was a mess and the output buffers had been removed, the other transistors run at very high bias current. All I could do was change the transistor package and add heat sinks to dissipate the heat while reducing the standing current somewhat. I added better regulation for the power supplies. The performance did improve considerably. Best of all, the preamp now didn't blow up anymore.
The Teflon PCB looked like it was designed for something different. The two external power supplies could have powered a small stereo amplifier each. It was a real PITA for the customer, so he was very happy to have but one external supply. Stereo separation was actually improved going to one supply, improving the regulators and taking care with grounds.
Factory DH-101 preamps performed better than this unit as received. So the money the customer spent upgrading this product was a total loss and resulted in something without any resale value.
This is exactly the kind of thing I have a massive problem with. When you accept money from another person to improve the performance of a piece of equipment, you should darn well do as you claim to. I feel it is irresponsible to charge a lot of money without any real proof of an improvement in performance. A bunch of people sitting around patting each other on the back doesn't count.
-Chris
I wasn't clear I guess.
The DH-101 I worked on was modified by John in this case, and he used a Teflon PCB. It was a mess and the output buffers had been removed, the other transistors run at very high bias current. All I could do was change the transistor package and add heat sinks to dissipate the heat while reducing the standing current somewhat. I added better regulation for the power supplies. The performance did improve considerably. Best of all, the preamp now didn't blow up anymore.
The Teflon PCB looked like it was designed for something different. The two external power supplies could have powered a small stereo amplifier each. It was a real PITA for the customer, so he was very happy to have but one external supply. Stereo separation was actually improved going to one supply, improving the regulators and taking care with grounds.
Factory DH-101 preamps performed better than this unit as received. So the money the customer spent upgrading this product was a total loss and resulted in something without any resale value.
This is exactly the kind of thing I have a massive problem with. When you accept money from another person to improve the performance of a piece of equipment, you should darn well do as you claim to. I feel it is irresponsible to charge a lot of money without any real proof of an improvement in performance. A bunch of people sitting around patting each other on the back doesn't count.
-Chris
Hi Bob,
I have observed that most "modification specialists" use products that were either advertised to be better (capacitors for example) and things that are expensive but don't apply to normal audio circuitry. The Rogers HF PCB material was one of those things and I suspect the entire reason it was used in the DH-101 I saw. Unless you are dealing with very high frequencies, or very high impedance circuits, these materials don't do a thing for you.
I use equipment to measure various aspects of capacitors (for example) at different frequencies to determine if they are better than a factory component. I have even found inductive metal film resistors! (surprised me) One very effective technique is to match diff pairs closely. Matching compliments as much as you reasonably can also helps, this is true for outputs and predrivers and driver transistors. That makes the positive and negative going transconductance more equal. The less the diff pair stage has to do to compensate for the circuit, the lower the distortion will generally be.
Coupling capacitors don't affect the sound much as long as they are reasonable and do not develop a voltage drop across them from the signal current. This should be obvious. As low a D-A you can reasonably find is probably a good idea.
Finally, any components you do use need to fit the mounting space. Parts that are too large cause all kinds of issues.
-Chris
I have observed that most "modification specialists" use products that were either advertised to be better (capacitors for example) and things that are expensive but don't apply to normal audio circuitry. The Rogers HF PCB material was one of those things and I suspect the entire reason it was used in the DH-101 I saw. Unless you are dealing with very high frequencies, or very high impedance circuits, these materials don't do a thing for you.
I use equipment to measure various aspects of capacitors (for example) at different frequencies to determine if they are better than a factory component. I have even found inductive metal film resistors! (surprised me) One very effective technique is to match diff pairs closely. Matching compliments as much as you reasonably can also helps, this is true for outputs and predrivers and driver transistors. That makes the positive and negative going transconductance more equal. The less the diff pair stage has to do to compensate for the circuit, the lower the distortion will generally be.
Coupling capacitors don't affect the sound much as long as they are reasonable and do not develop a voltage drop across them from the signal current. This should be obvious. As low a D-A you can reasonably find is probably a good idea.
Finally, any components you do use need to fit the mounting space. Parts that are too large cause all kinds of issues.
-Chris