Hafler DH-200/220 Mods

Hi Rick,
Dual holes for stuffing options are fine, familiar with the practice. I'm also familiar with the hopeless bypassing of coupling capacitors. That practice is ill conceived. 1 uF or 10 uF? Well, 10uF is a bit overkill depending on the input impedance. A choice? DIstortion and "sound quality" is insensitive to coupling capacitor type unless you use defective or really horrible types. I haven't tested tantalum caps, but who would use them these days? I'm not even sure if you could hear high K or tantalum to be honest.

Stuffing options are fine unless the resulting PCB layout (pattern) causes problems. Otherwise, you don't have any concerns. I would have used holes for the leaded transistors and adapters for smt parts. But then I like clean, simple layouts and come from a service background where an adapter PCB is a real opinion.

Component selection. Those large axial types hanging up towards the sky are probably not a good choice.
 
Agreed. Use leaded through-hole resistors. It will simplify your life and will improve reliability in the long run. Also, many MELFs are inductive. It may be meaningless here but maybe not depending on the location.
Also, why do you seek an alternative for the capacitors? Is it because you can't get the specified component?
The circuit shows a 10 uF electrolytic bypassed with a 1 uF polyester type.
Would you prefer a polypropylene cap instead of the specified polyester?
I must admit that I have never used a MELF resistor. That doesn't mean they are bad, just that I think good-quality metal film resistors are more than adequate for the highest level of audio quality. Although not relevant to this build, I would be cautious about using surface mount resistors in critical locations, like the feedback path and some other places in the signal path. Just my old fashioned cautiousness.

I generally hate electrolytics in the signal path, and just bypass them out of good measure, even though arguments that bypassing them doesn't make a difference are not unreasonable. There are also many who say it makes no difference whether one uses polyester capacitors versus polypropylene (or even polystyrene). There are reasonable reasons given on either side of that barbed-wire fence, but I usually stick with the higher-quality capacitors out of an abundance of caution. The cost difference is minimal and some audiophiles might shy away from my designs if I used polyester capacitors in the signal path.

Cheers,
Bob
 
  • Like
Reactions: anatech
Hi Bob,

I used the MELF on Sam Groners LNA. They are fine to use, highly recommended, but are more expensive. It's not the point of using a MELF, its the way he applied it, looks as if he tack/blob soldered a small piece of lead wire onto the metal face. Kinda defeats the purpose of a leadless component. Looks? to be 0207 because 0204 are too difficult to do. Is it a reliable connection? They (0207) are so big that he has to mount the adjacent ones underneath. But right beside them he is using other small recommended types, SFR16s,MF0204, MBB0204? as in the BOM, which are all good choices. It's hard to understand why. It's a lot of work for no gain, and risky too, so that why I says its kinda crazy but theatrical. I hope it works in the end because its a lot of effort and $.

I hope folks have the KSA1381 and KSC3503 devices in hand now that they are EOL? I'd like to try out TTA004B,TTC004B as drivers, and VAS to see how they compare, that and the KEC KTA1381 and KTC3503 devices that Profusion is going to carry to save the day.

I had contemplated doing the DH-220C layout in smt, because its a lot of work to stuff and solder them up. This is where you could use a MELF if you too.

The coupling capacitor debate I do even bother with, but it is entertaining 🙂
I do like the emerald green Nichicon UES, goes good with the Kemet green stack polys and the red Wima PP's kinda goes with the season..

Rick
 
The KSA1381 and KSC3503 have been wonderful devices to work with. I could be wrong, but I think they were designed for use in CRT-based high definition TVs, where fast, high-voltage transistors were needed. I am not familiar with the TTA004B and TTC004B devices, but they may be worth a look. In the meantime, I have ordered more KSA1381 and KSC3503 transistors.

Cheers,
Bob
 
I don't trust smt devices with higher voltages across them. Even if they don't fail, you have leakage concerns (on the PCB) and heat generation. I really do not like avoidable heat.

For resistors, you have voltage gradient non-linearity of varying degrees, so I use 1 watt resistors in the feedback path, or series connected resistors. Also choose low voltage coefficient parts. As for capacitors, I'll use the better ones simply due to caution. I have tested them and bypassing larger caps with whatever you scare up won't change anything. Not unless you put a leaky, terrible part across an existing capacitor. What they do is mess up a layout, or hang across a part after someone adds them. Those get pulled first in a repair.

Think about it. You have a large electrolytic with whatever issues it may possibly have. The circuit impedance is what? So that little part is going to reverse DA or leakage? Nope. Damage done if you can prove the electrolytic did any damage to start with. Until you generate a signal voltage across the capacitor, it won't affect the sound unless it upsets circuit conditions. I proved this over the decades using progressively better equipment. So theory matches practical experience. Hmmm.
 
I don't trust smt devices with higher voltages across them.
get over it, it's done all the time at much higher voltages than an audio amp sees. The SOT-223 is a great package to use up to a few watts.

Bob, I'll sample some KTA1381 and KTC3503 devices to try out and send you some to try out as well.
TTC004B and TTA004B are very economical Toshiba parts, not to be used in the DH-500C as they are 160V parts. TTA006,TTC011 are higher voltages.
 

Attachments

Last edited:
Hi Rick,
Nothing to get over. I've seen tracking in audio equipment using leaded parts. It happens, reducing distances will encourage it more. Surface mount components often has shorted distances between leads. Of course, that is the entire point of a surface mount part.

I am not saying surface mount construction is bad. It easily can be, and temperature rise with leaded parts is less than surface mount parts.

I've also seen first hand what happens when you get many warm parts closer together. Reliability drops, no shock there as you increase the running temperature of devices. It may not affect failure rates much a few years out, but after 10 or so it really does. 10 years will go by quickly for equipment owners. I'm working on 40 + year old equipment capable of high power that don't suffer from heat related issues, but newer stuff that is trashed, not reliably repairable considerably newer. Even the PCBs become unreliable in those instances.

Unless you pay attention to heat and voltage gradients, you do sentence the equipment owner to problems in the future. That is a fact.
 
Surface mount components often has shorted distances between leads.
I purposely mentioned TO-223, lead spacing is 2.4mm, a TO-92 is 1.27mm.
Which do you think is better for your high voltage and leakage concerns?
A to-220 has 2.54mm lead spacings
I use the BYV34X-600 diode, I am amazed that its a 600V/20A rated part (SOT186A (3-lead TO-220F)

Your new Toyota is full of smt, good luck to you, I hope you are wrong in your smt concerns, your car is ultra reliable, you are not dead on the side of the road with a big bill for that $0.10 Chinese diode that cracked. With that smt talk down, I'd get extended warranty too 🙂
Did that NASA space probe that flew closet to the sun have any smt? 🙂
Anyways taking the Hafler thread into the reliability camp, for another thread and time.
Maybe I'll design the DH-220C in smt and you can beat the hell out of it in your reliability testing. See what a MELF and SOT-223 can handle.
 
Last edited:
The KSA1381 and KSC3503 have been wonderful devices to work with. I could be wrong, but I think they were designed for use in CRT-based high definition TVs, where fast, high-voltage transistors were needed.
Exactly. That's the point why we have them - and why they'll soon be EOL.
Until you generate a signal voltage across the capacitor, it won't affect the sound unless it upsets circuit conditions. I proved this over the decades using progressively better equipment. So theory matches practical experience. Hmmm.
Well said. Nothing to add. The only applications where inferior dielectrics might result in poorer performance, are filters, e. g. RIAA equalisation networks etc. But even this still needs to be proven.

Best regards!
 
  • Like
Reactions: anatech
Hi Rick,
The brand new Toyota (Matsushita) radio already failed within months, took 3 months for a replacement. The trunk sometimes opens without command, once while I was standing beside the car.

Electronic subassemblies commonly fail in cars, more in newer cars than ever before. Two different friends repair them, one makes new modules for the ones discontinued (nice way to force a new car eh?). This is with toughened parts designed for automotive vibration and temperature swings plus power supply glitches. Many have conformal coating. Our new university educated designers believe the max ratings in data sheets and design accordingly. That's something you or I would never do because we know better. My brother owns some large used car dealerships, on the web and both in the US and Canada, several provinces. He will not touch EV or Hybrids due to failures and nightmares getting them fixed - and the very high costs involved. I've repaired some automotive electronics since the 1980's for him, so i have a little familiarity. I also did automotive audio for several top end brands (McIntosh, Nakamichi, Marantz, Denon, Alpine and a few others). I've seen the technology change from the first Fosgates and MGT amps and the failure rates and types of failures.

Rick, I know the part you selected had wider collector to other terminal spacing. But it is also designed for higher powers over the normal package, so you didn't prove anything by example. TO-220 and larger packages normally have the collector or drain lead bent away at higher voltages, increasing the spacing. They do the same for TO-92, but you can't with smt signal packages, can you? Any guesses as to why they increase terminal spacing with leaded parts?

I also lay out PCBs with smt transistors. I pay careful attention to power density and voltage ratings. However, most folks don't and create the very situations I mentioned. Early in time, I discovered my Philips PM3070 scope died 4 times due to smt transistors mounted too close, running too hot. They repaired it under warranty each time, then offered their new "Combiscope" as a direct warranty replacement. I measured a 20°C temperature rise between inlet and outlet air temperature !!!! I traced that to one huge IC without a heat sink. Needless to say, I kept the PM3070. When it dies again (it's only on when I use it), I know I'll be replacing a bunch of smt parts and maybe the darned PCB again since it goes chocolate brown. What do you think the failures are due to?

All I am saying is that smt encourages higher circuit density. The packages are not as good at getting rid of heat and that heat goes pretty much directly into the PCB. So you concentrate higher heat in a smaller area and guess what happens? My only comment would be that you need to space things out, and use larger smt packages and/or heat sinks to avoid higher operating temperatures. That is all I am saying.

I don't have a hate on for smt as long as you use it properly. There is a time and place for tiny smt packages, and other times where you really should use the larger, leaded parts. That is if you want reliability lasting over 5 years. Most people expect their audio amps and preamps to last over 20 years without failure. Pretty reasonable given we had that.
 
I don't trust smt devices with higher voltages across them. Even if they don't fail, you have leakage concerns (on the PCB) and heat generation. I really do not like avoidable heat.

For resistors, you have voltage gradient non-linearity of varying degrees, so I use 1 watt resistors in the feedback path, or series connected resistors. Also choose low voltage coefficient parts. As for capacitors, I'll use the better ones simply due to caution. I have tested them and bypassing larger caps with whatever you scare up won't change anything. Not unless you put a leaky, terrible part across an existing capacitor. What they do is mess up a layout, or hang across a part after someone adds them. Those get pulled first in a repair.

Think about it. You have a large electrolytic with whatever issues it may possibly have. The circuit impedance is what? So that little part is going to reverse DA or leakage? Nope. Damage done if you can prove the electrolytic did any damage to start with. Until you generate a signal voltage across the capacitor, it won't affect the sound unless it upsets circuit conditions. I proved this over the decades using progressively better equipment. So theory matches practical experience. Hmmm.
I've seen your argument about the unimprovement of shunting electrolytics many times, and I have often thought about it myself. But strictly "meter-reader" objective arguments don't always cut it with many audiophiles and listening tests (many of which, admittedly, are hopelessly flawed). This is an age-old debate not worth having here. I just err on the side of doing what some have done to achieve better sound. I measure like crazy, but don't let that rule out some things that some think come out of listening tests. I don't dismiss things like the work on "capacitor sounds" by Walt Jung and many others over the years. It is extremely difficult to justify using polypropylene capacitors over polyesyer capacitors based merely on measurements of the performance of equipment using one or the other.

For many decades I have been curious of, and deeply interested in, why some audio equipment sounds different or not as good as some other when it is hard to correlate that with objective measurements. Here is my middle-of-the road take: Half of the stuff that is promoted in high-end audio is snake oil; but we often don't know which half. Speaker cables are but one of these issues where there is legitimate controversy. Very expensive power cables are probably even worse. This does not mean that clean power free of EMI etc is not important; rather it means that most of those extremely expensive power cables likely don't do squat to remedy issues like EMI etc. If you are talking about expensive power conditioners with great filtering and maybe even isolation transformers, there is more credibility to that.

I agree completely about keeping signal voltages low across resistors because of voltage-dependent nonlinearities. Also about keeping signal-dependent temperature swings low. Use of feedback resistors of higher wattage, and often in series is usually a good idea.

Cheers,
Bob
 
get over it, it's done all the time at much higher voltages than an audio amp sees. The SOT-223 is a great package to use up to a few watts.

Bob, I'll sample some KTA1381 and KTC3503 devices to try out and send you some to try out as well.
TTC004B and TTA004B are very economical Toshiba parts, not to be used in the DH-500C as they are 160V parts. TTA006,TTC011 are higher voltages.
Thanks, Rick. I'll take a look at these spec sheets and let you know what I find.

Cheers,
Bob
 
Hi Bob,
That's the problem. Most "meter readers" listen carefully. Also, they are not meters, they are spectrum displays. Folks who are loudly "subjectionists" ignore the fact that "meter readers" are also subjective with readings to help guide and explain. We also employ subjective type observers to gauge their response. Nakamichi and Cyrus both had careful listening rooms to assess their products, others as well I'm sure. Those companies also had full labs with very good equipment.

I have always had others judge my work or changes. There is zero chance any of us could be objective on our own work. I tend to be overly critical for example. My observations have been ongoing since the 1970's when I worked in a high end audio store. I soon hung a spectrum analyzer on the output of various THD meters. That revealed a huge amount of information. The HP 3580A was amazing, followed by several other types, an HP 35665A, 3585A and finally an RTX 6001. Each one revealing more information and largely confirming hunches. I would love an AP555 series, or Keysight U8903B, I cannot justify the price compared to the excellent little RTX. Especially since I already have a full bench of pricy gear.

Differences in equipment has little to do with capacitors and resistors with the better brands. The parts they use do not limit performance. They just don't use parts known to cause distortion, they leave that to cheaper brands and "designers" who like to "voice" their equipment. What seems to really matter is PCB layout and common (ground) paths. Over many years and observing people's reactions, the best sounding equipment has the lowest distortion and background noise. Of course, an amplifier has to dump current into more difficult loads without excessive increased distortion. All within reason.

What I found very interesting lately is replacing a Marantz 300DC (they sound "edgy" stock) I had modified with a Yamaha PC2002 I repaired and cleaned up. They are driving a pair of PSB Stratus Gold speakers, not an easy load. They measured closely on the RTX under load. Guess what? They sound the same in operation on those speakers. My goal was higher power, and that was achieved without sacrificing sound quality. Now, there is a slightly better amplifier I had for 3 weeks. I would love one. The Bryston 4B cubed, and I didn't like Bryston products historically.

So proper measured performance correlates well with observations from the average person. The big problems come when an audiophile has industry nonsense in the back of their heads, but give them a while and they always see things as they really are. That is unless they have some investment in a particular viewpoint. Expectation bias is your biggest problem.

Anyway, I have spent my life in audio. I don't expect people fresh or with much less experience to be able to figure things out with what you can read on the web confusing issues. If you could isolate them from advertising and just let them play with the gear, they could figure all this out much more quickly.

I have been asked to study power cables, speaker cables, RCA cables (and XLR) and AC power filters and regenerators since the early 1980's. Most is only effective with poorly designed audio equipment, or special situations. RCA cables with high impedance equipment for example, shielding and DA affect the sound, capacitance can when it is extreme. That would be tube equipment without buffers or volume controls in a box (passive preamps, but you shouldn't use that term).

Power cables only matter if your internal power supply is defective, poorly designed. No audio equipment manufacturer of good equipment would ever allow a power cord to impair their equipment. Power amps should always, without exception, be plugged directly into the AC wall outlet. A direct run to the panel on the same phase is best for higher power amps. A filter may possible help some signal equipment (preamps, signal sources), but it shouldn't affect anything if they are quality pieces, or unless you have severe AC line quality issues.

Speaker wire. Sigh ... I have see some cause instability in some amps. The best? High quality copper "zip cord" style of reasonable size. 16 ga is normally more than enough, your terminations have the highest resistance, not the wire. Then if you have a passive crossover, the crossover determines everything unless you go silly small with the wire.

People tend to focus on small things in situations that don't reflect the use. RF effects for example. Balanced. Great, now you installed two adapter circuits in series with the signal. Balanced lines can solve specific problems, but like everything in engineering it is a balance between solving and issue and other issues that come up. You go for the best outcome. Long distance runs with high electrical noise? Balanced. Otherwise you are further ahead going single ended, the equipment is normally single ended anyway. Balanced gear is 1.414 x noisier.

Focus on the engineering. Consider advantages and disadvantages and come up with the best solution. No one thing is best in every situation.

My 2 cents.
 
I purposely mentioned TO-223, lead spacing is 2.4mm, a TO-92 is 1.27mm.
Which do you think is better for your high voltage and leakage concerns?
A to-220 has 2.54mm lead spacings
I use the BYV34X-600 diode, I am amazed that its a 600V/20A rated part (SOT186A (3-lead TO-220F)

Your new Toyota is full of smt, good luck to you, I hope you are wrong in your smt concerns, your car is ultra reliable, you are not dead on the side of the road with a big bill for that $0.10 Chinese diode that cracked. With that smt talk down, I'd get extended warranty too 🙂
Did that NASA space probe that flew closet to the sun have any smt? 🙂
Anyways taking the Hafler thread into the reliability camp, for another thread and time.
Maybe I'll design the DH-220C in smt and you can beat the hell out of it in your reliability testing. See what a MELF and SOT-223 can handle.
I think one of the biggest reliability problems today is the mandated use of lead-free solder in commercial equipment, thanks to the EU ROHS mandate. The joints often look like cold solder joints even when they are good. The significantly higher temperatures needed for reflow may also be a problem in some cases. Heaven help you if you have a big BGA chip where you cannot inspect the balls. Good that DIYers can still use leaded solder.

Cheers,
Bob
 
Exactly. That's the point why we have them - and why they'll soon be EOL.

Well said. Nothing to add. The only applications where inferior dielectrics might result in poorer performance, are filters, e. g. RIAA equalisation networks etc. But even this still needs to be proven.

Best regards!
The use of the word "proven" in audiophile circles can be problematic 🙂.

Regarding phono preamps, the traditional RIAA phono preamp feedback circuit often involves a shunt electrolytic capacitoe in the feedback circuit. This is definitely in the signal path, and is required to have a low impedance throughout the audio band. I like to avoid using an electrolytic there by not using a shunt capacitor there at all to reduce the gain to unity. Instead, I prefer to use a DC servo with an audio-grade op amp serving as the integrator for the DC servo. I also use an audio-grade polypropylene integrator capacitor. A DC servo is definitely in the signal path.

Cheers,
Bob
 
Hi Bob,
That electrolytic capacitor isn't always audible. That and noise from the servo may be a worse problem depending on how you do it.

Hi Jack,
Thanks for the link. This underscores what I learned in the calibration lab and in designing instruments. He's right, this is a kind of secret only some really understand.
 
Hi Bob,
That's the problem. Most "meter readers" listen carefully. Also, they are not meters, they are spectrum displays. Folks who are loudly "subjectionists" ignore the fact that "meter readers" are also subjective with readings to help guide and explain. We also employ subjective type observers to gauge their response. Nakamichi and Cyrus both had careful listening rooms to assess their products, others as well I'm sure. Those companies also had full labs with very good equipment.

I have always had others judge my work or changes. There is zero chance any of us could be objective on our own work. I tend to be overly critical for example. My observations have been ongoing since the 1970's when I worked in a high end audio store. I soon hung a spectrum analyzer on the output of various THD meters. That revealed a huge amount of information. The HP 3580A was amazing, followed by several other types, an HP 35665A, 3585A and finally an RTX 6001. Each one revealing more information and largely confirming hunches. I would love an AP555 series, or Keysight U8903B, I cannot justify the price compared to the excellent little RTX. Especially since I already have a full bench of pricy gear.

Differences in equipment has little to do with capacitors and resistors with the better brands. The parts they use do not limit performance. They just don't use parts known to cause distortion, they leave that to cheaper brands and "designers" who like to "voice" their equipment. What seems to really matter is PCB layout and common (ground) paths. Over many years and observing people's reactions, the best sounding equipment has the lowest distortion and background noise. Of course, an amplifier has to dump current into more difficult loads without excessive increased distortion. All within reason.

What I found very interesting lately is replacing a Marantz 300DC (they sound "edgy" stock) I had modified with a Yamaha PC2002 I repaired and cleaned up. They are driving a pair of PSB Stratus Gold speakers, not an easy load. They measured closely on the RTX under load. Guess what? They sound the same in operation on those speakers. My goal was higher power, and that was achieved without sacrificing sound quality. Now, there is a slightly better amplifier I had for 3 weeks. I would love one. The Bryston 4B cubed, and I didn't like Bryston products historically.

So proper measured performance correlates well with observations from the average person. The big problems come when an audiophile has industry nonsense in the back of their heads, but give them a while and they always see things as they really are. That is unless they have some investment in a particular viewpoint. Expectation bias is your biggest problem.

Anyway, I have spent my life in audio. I don't expect people fresh or with much less experience to be able to figure things out with what you can read on the web confusing issues. If you could isolate them from advertising and just let them play with the gear, they could figure all this out much more quickly.

I have been asked to study power cables, speaker cables, RCA cables (and XLR) and AC power filters and regenerators since the early 1980's. Most is only effective with poorly designed audio equipment, or special situations. RCA cables with high impedance equipment for example, shielding and DA affect the sound, capacitance can when it is extreme. That would be tube equipment without buffers or volume controls in a box (passive preamps, but you shouldn't use that term).

Power cables only matter if your internal power supply is defective, poorly designed. No audio equipment manufacturer of good equipment would ever allow a power cord to impair their equipment. Power amps should always, without exception, be plugged directly into the AC wall outlet. A direct run to the panel on the same phase is best for higher power amps. A filter may possible help some signal equipment (preamps, signal sources), but it shouldn't affect anything if they are quality pieces, or unless you have severe AC line quality issues.

Speaker wire. Sigh ... I have see some cause instability in some amps. The best? High quality copper "zip cord" style of reasonable size. 16 ga is normally more than enough, your terminations have the highest resistance, not the wire. Then if you have a passive crossover, the crossover determines everything unless you go silly small with the wire.

People tend to focus on small things in situations that don't reflect the use. RF effects for example. Balanced. Great, now you installed two adapter circuits in series with the signal. Balanced lines can solve specific problems, but like everything in engineering it is a balance between solving and issue and other issues that come up. You go for the best outcome. Long distance runs with high electrical noise? Balanced. Otherwise you are further ahead going single ended, the equipment is normally single ended anyway. Balanced gear is 1.414 x noisier.

Focus on the engineering. Consider advantages and disadvantages and come up with the best solution. No one thing is best in every situation.

My 2 cents.
Just about everything you say is right on target. Meter readers + listeners = good audio. One thing I believe in is that if you design for good numbers that are thoroughly made and correctly interpreted, you leave less room for other issues that degrade sound performance that you may not be aware of. One time, along time ago, I was working on an amplifier and it had THD at higher frequencies that was somewhere in the 0.01 to 0.05% range. Many people would say that is OK. There was one problem: I knew that design and knew that its THD should not be that high. Looking more carefully, the amplifier had a local high-frequency parasitic oscillation on the back porch of a high-amplitude sine wave occurring in the output stage. Once I fixed that, the distortion went way down.

I've used an HP3580A for many decades. A wonderful instrument. Before I was able to get my hands on a used one, I used to borrow and take one home on occasion from Bell Labs where I worked. Your mention of using the 3580A starting a long time ago probably puts you in the same age group as me 🙂. I built several tube amplifiers when I was a teenager before I really knew what I was doing. My favorite AC voltmeter is the HP 400EL.

I have also never had an AP, so I built one myself in the early 1980s, and have made a habit of looking at its residual output. I have a stock HP339A, but mine has better performance. Lately I have been using the QuantAsylum QA403, and it has astounding performance for only $600. I highly recommend it. Just recently I wrote a tutorial on using it for testing power amplifiers. I used the BC-1 as the amplifier to test as an example.

Cheers,
Bob
 
  • Like
Reactions: anatech
Hi Bob,
That electrolytic capacitor isn't always audible. That and noise from the servo may be a worse problem depending on how you do it.

Hi Jack,
Thanks for the link. This underscores what I learned in the calibration lab and in designing instruments. He's right, this is a kind of secret only some really understand.
Not in my DC servos. One just needs to understand how to design them properly. I discuss them at length in my amplifier book.