"refurbishing" a Hafler DH110 Preamp

Talon

Member
2014-08-01 11:54 pm
My first preamp was a DH100, and I regret ever selling it. I'd always wanted a DH110, but by the time I'd acquired this one I had so much other gear to use that I never bothered to give it any work to do. It went into storage.

When I did hook it up recently, I got some really disturbing results out of it. Overall the sound wasn't good, and I know these pre-Rockford Hafler units can be very nice sounding. The volume pot in particular was very bad. At around the 1/3 mark it had a sharp transition between barely audible and listenable, and it made cracks loud enough there that I thought I'd better power the thing down, and give it away.

Or fix it.

I pulled open the cover, and started looking over the insides:

- No debris, no dust even, so the PO obviously did some preliminary work
- Two of the eight BJTs on the board that wore heat sinks were missing them
- no electrolytic caps have burst or leaked, but a few of them exhibited a little bit of "pillowing" on top, where the "Mercedes logo" relief cut had a little hill in each of the three sections
- other than a tiny bit of what's probably cleaning residue near some clusters of BJTs (and nowhere near any caps), this board is pristine


My questions so far:

- I'm pretty sure the pots need to be cleaned. Especially that volume pot. I'm new to this and have never done that work. I have CRC/QD contact cleaner, MaxPro contact cleaner, and Deoxit F5 on hand. I'm sure that whichever one of these I try to use, it will be the wrong one, I'll put it in the wrong spot, it'll destroy the pot, and when I come back here to report it, everyone will say "why did you use that? You should have used (x product)!" so I figure I'd ask here first - what should I use, and how?

- Hafler used a frame-type transformer mounted diagonally in the back corner, so I can't convert to an IEC plug without major surgery (no room). Has anyone here done an IEC retrofit on the DH110, or at least upgraded to a better 3 wire cord?

- I'd like to source the caps myself. If I replace just a few caps, I should replace all 24 of them. The Hafler assembly manual parts list is riddled with typos and is not brain-friendly (the BJT list is especially badly done - whomever did it freely mixed up the letter I with the number 1). Would anyone here have a more concise list available?

- If I turn out to be not smart enough to figure out the cap values, is there anything wrong with going to "that guy" on eBay and buying the complete cap package for $43?

- What brand of capacitors are recommended? I am open-minded to both the arguments against esoteric caps as well as the arguments in favor of them. I think both have good points, but there's a law of diminishing returns. Silver sounds better than tin, but I can't tell the difference between gold and silver -and apparently some people can- is what I'm saying. So what "audiophile-grade" capacitors are a good value, if there is such a thing? I'm also open to buying a lot for a volume discount, if that helps.

- Where do I find the goofy little stamped copper heat sinks for these BJTs? They're about 2/3 the size of postage stamps and made to fit TO-92 packages. They're probably pretty important.
 

Talon

Member
2014-08-01 11:54 pm
This should be fun.. I imported the list of cap values, and it looks like I have 41 caps to replace. The misspellings are irritating:

10 mF, polypropykne, 5Ov
120 pF, polypropylene, 50v
220 pF, polypropykne. 50v
10nF polypropylene, 5Ov

22 mfd, NP ekctmlytic, 5Ov
220 mfd, NP ekctmlytic, IOv
220 mfd, NP ekctrolytic, 10v
I5 nF, polycarbonate, 5Ov
I5 nF, polycarbonatc, 5Ov

25 nF I%, poiypropykne, 50v
6.8 nF, 1%. polypropylene, 5Ov
1000 mfd, NP ekctrolytic, 6v
1.5 nF, polypropykne, 50v

IO nF, polypropylene, 50v

22 mfd, NP ekctrolytic, 50v
470 mfd, low ESR electrolytic,
25v 470 mfd, low ESR electrolytic,
25v 470 nF, polypropylene, 50v

68 pF, polypropylene, 50v
10 nF, polypropylene, 50v

22 mfd, NP electrolytic, 5Ov

470 mfd, low ESR electrolytic,
25v 470 mfd, low ESR electrolytic,
25v I .5 nF, polypropylene, 50v

I5 nF, polycarbonate, 50v

47 nF, polycarbonate, 5Ov

470 nF, polycarbonate, 5Ov

4.7 mfd, NP ekctrolytic, 25v

47 mfd, NP ekctrolytic. l6v

100 nF, polycarbonate. 50v
1000 mfd, electrolytic, 50v
1000 mfd, electrolytic, 50v
22 mfd, ekctmlytic, 50v

100 nF, polycarbonate, 50v
22 mfd, ekctrolytic, 50 v

100 nF. polycarbonate, 50v
47nF, polycarbonate, 50v

4.7 mfd, NP ekctmlytic, 25v
4.7 mfd, NP ekctrolytic. 25v
22 mfd, electrolytic, 50v
 

Talon

Member
2014-08-01 11:54 pm
Start with the electrolytics. The film caps rarely go bad. After that you can experiment if you wish with Boutique caps in the signal path like Russian Teflon....

Well thats good news, that gets me down to 19 caps:


4.7 mfd, NP electrolytic, 25v (3)

22 mfd, electrolytic, 50v (3)
22 mfd, NP electrolytic, 50v (3)

47 mfd, NP electrolytic. 16v

220 mfd, NP electrolytic, 10v (2)

470 mfd, low ESR electrolytic, 25v (4)

1000 mfd, NP electrolytic, 6v

1000 mfd, electrolytic, 50v (2)


I'm a bit curious about the 47mfd cap and the 1000mfd, due to the odd voltages. One is 16v, the other is 6v? I'm reasonably sure this board runs at 16v, if its anything like my IRIS preamp, so that 6v part is probably another one of the long list of typos. Of course I'll verify the numbers based on the parts installed now.


So I should look all these electrolytics up as Nichicon parts, right? Is there anything better out there?
 

Talon

Member
2014-08-01 11:54 pm
Trying to shop Mouser and Digithief is like trying to find a needle in a needle stack. Their "search" system leaves a lot to be desired. I'm tempted to shop with one of the eBay guys just to save the time, but I've already spotted weasel words in their ads that lead me to believe I won't be getting what I thought I'd get.
 
Your hunch is quite likely to be correct - better not trust dodgy eBay sellers.

Nichicon is good for nonpolars. Otherwise, Panasonics are very good caps as well. Wouldn't be surprised if the standard FM or FC series would already qualify as "low ESR" by the standards of the day, let alone FR.

Take measurements of original caps and see what the voltage rating comes out as if you go with same capacitance and physical size. It's usually a good idea to uprate voltage by 1 step at least. Any bump above 63-100 V (or about 3x sustained voltage) is not likely to be worth it though.

6 V may be a typo indeed, as the standard value would be 6.3 V. Usual steps are 10 - 16 - 25 - 35 - (40 -) 50 - 63 - (80 - ) 100.
 
Trying to shop Mouser and Digithief is like trying to find a needle in a needle stack. <snip>

Having bought many thousands of caps over the years from several vendors I can tell you that DigiKey is one
of the best if not THE best for searching out parts. When I buy caps the physical size is foremost. Voltage
at least what the original was and I usually go higher as the caps have been getting smaller. The ripple current
is the determining factor as it relates directly to ESR. Usually I look for max hours at 105° but sometimes 85°
caps are fine. Panasonic FR FT and FC are preferred but I've bought a fair number of Nichicons, Rubycons and
United Chemicons. Try again, it's not hard.
 

Talon

Member
2014-08-01 11:54 pm
Having bought many thousands of caps over the years from several vendors I can tell you that DigiKey is one
of the best if not THE best for searching out parts. When I buy caps the physical size is foremost. Voltage
at least what the original was and I usually go higher as the caps have been getting smaller. The ripple current
is the determining factor as it relates directly to ESR. Usually I look for max hours at 105° but sometimes 85°
caps are fine. Panasonic FR FT and FC are preferred but I've bought a fair number of Nichicons, Rubycons and
United Chemicons. Try again, it's not hard.

Well, I wouldn't say its hard to do, its just time consuming and full of rabbit holes.

But by your advice, I gave it another try, and I may have actually gotten somewhere this time. I looked for a 1000uF radial, 50v, +/- 20 (not happy about that, but they had nothing in better tolerance), 10000 @ 105c, FR series Panasonic. I got two results and they look to be identical with the difference being the packaging. Two different Digikey part numbers resolve to a single Panasonic part, EEU-FR1H102B. If I buy them "tape and box", they're 99 cents each.

But is that the right part for the job?

The more I talk to people the more I learn, and then it seems I'm farther away from making a decision than I was the day before. You're the first person to bring ripple current up in any discussion I've had about caps. This part states [email protected] 120Hz, [email protected] Does that work for my application? I have no way to know. Also, you mention you shop cap size first. The only observations I've seen from other people on that topic are how you can fit bigger/more caps in the same spot as a vintage cap. What exactly are you concerned about with cap size? What is your goal?
 
Well, I wouldn't say its hard to do, its just time consuming and full of rabbit holes.
<snip>

First of all, 'lytic caps USED to be +80/-20 % tolerance so they are quite a bit better on that score. Reason I
mention size first is there is no point buying something that won't fit. Since I service many different units I
try to buy parts to be more 'universal' so if the unit uses 100uF at 16V and I can now get the same size at
25V or 35V I'll buy the higher Voltage to reduce inventory.

The ripple current directly relates to internal heating. NONE of the equipment ever gets to 105° C but the cap
can get hot from the charging/discharging currents. That's why caps bulge. The lower ESR (higher ripple
current) reduces that heating so it runs longer before it fails again. This is extra important in switching
power supplies. The linear supply in the Hafler would be OK with almost any cap of the right capacitance
and Voltage. The higher ripple current will likely cause a higher inrush current when turned on. The transformer
and diode bridge will be stressed a bit more for a few milliseconds but I promise you won't have a problem with it.
With a little expectation bias you can convince yourself it now sounds better. If the caps were bad enough it
may very well sound better.

 
This should be fun.. I imported the list of cap values, and it looks like I have 41 caps to replace. The misspellings are irritating:

10 mF, polypropykne, 5Ov
120 pF, polypropylene, 50v
220 pF, polypropykne. 50v
10nF polypropylene, 5Ov

22 mfd, NP ekctmlytic, 5Ov
220 mfd, NP ekctmlytic, IOv
220 mfd, NP ekctrolytic, 10v
I5 nF, polycarbonate, 5Ov
I5 nF, polycarbonatc, 5Ov

25 nF I%, poiypropykne, 50v
6.8 nF, 1%. polypropylene, 5Ov
1000 mfd, NP ekctrolytic, 6v
1.5 nF, polypropykne, 50v

IO nF, polypropylene, 50v

22 mfd, NP ekctrolytic, 50v
470 mfd, low ESR electrolytic,
25v 470 mfd, low ESR electrolytic,
25v 470 nF, polypropylene, 50v

68 pF, polypropylene, 50v
10 nF, polypropylene, 50v

22 mfd, NP electrolytic, 5Ov

470 mfd, low ESR electrolytic,
25v 470 mfd, low ESR electrolytic,
25v I .5 nF, polypropylene, 50v

I5 nF, polycarbonate, 50v

47 nF, polycarbonate, 5Ov

470 nF, polycarbonate, 5Ov

4.7 mfd, NP ekctrolytic, 25v

47 mfd, NP ekctrolytic. l6v

100 nF, polycarbonate. 50v
1000 mfd, electrolytic, 50v
1000 mfd, electrolytic, 50v
22 mfd, ekctmlytic, 50v

100 nF, polycarbonate, 50v
22 mfd, ekctrolytic, 50 v

100 nF. polycarbonate, 50v
47nF, polycarbonate, 50v

4.7 mfd, NP ekctmlytic, 25v
4.7 mfd, NP ekctrolytic. 25v
22 mfd, electrolytic, 50v
So many typos in that list...
Those typos are typical from scanning a printed sheet (which is an image, not a text file) and converting it using OCR software.

Scanner was lazy, you always need to correct such rubbish output.

Our brain has very good processing built-in which allows us to understand:

t0m0rr0w w3 w1LL b3 wr1t1n6 4 n3w v3rs10n 0f th15 t3xt ,

while a computer will be baffled to distinguish between scanned capital i "I" and "l" and "1" in many cases, and so on.
 
...plus puzzling unit denominations. What for heaven's sake does »mfd« mean in a worldwide SI world???
We all know: The capacitance unit is the Farad, abbreviated F. 1 (one) F is a rather large capacitance, only met by some storage or buffer capacitors (»Gold Caps«). Hence in practical usage we need smaller units, such as:

1 mF = 10-3 F
1 µF = 10-6 F
1 nF = 10-9 F (astonishingly this can be found in the list above)
1 pF = 10-12 F

Best regards!
 
...plus puzzling unit denominations. What for heaven's sake does »mfd« mean in a worldwide SI world???
We all know: The capacitance unit is the Farad, abbreviated F. 1 (one) F is a rather large capacitance, only met by some storage or buffer capacitors (»Gold Caps«). Hence in practical usage we need smaller units, such as:

1 mF = 10-3 F
1 µF = 10-6 F
1 nF = 10-9 F (astonishingly this can be found in the list above)
1 pF = 10-12 F

Best regards!

That list is likely form the early '80s or late '70s. Nobody worried about the "SI" world and everyone
(at least on the US) knew mfd was a microfarad. The other one that would REALLY bug you is micro micro farad
(mmf) now known as a pico farad. I'm at retirement age and still not comfortable with nano farads.

 

Talon

Member
2014-08-01 11:54 pm
Those typos are typical from scanning a printed sheet (which is an image, not a text file) and converting it using OCR software.

Scanner was lazy, you always need to correct such rubbish output.

Our brain has very good processing built-in which allows us to understand:

t0m0rr0w w3 w1LL b3 wr1t1n6 4 n3w v3rs10n 0f th15 t3xt ,

while a computer will be baffled to distinguish between scanned capital i "I" and "l" and "1" in many cases, and so on.

Understood, but I typed that list letter-for-letter from the service manual, not a from an OCR scan. Those typos were original Hafler work. They must have put all the effort into designing and building good components and not into technical docs.
 

Talon

Member
2014-08-01 11:54 pm
First of all, 'lytic caps USED to be +80/-20 % tolerance so they are quite a bit better on that score. Reason I
mention size first is there is no point buying something that won't fit. Since I service many different units I
try to buy parts to be more 'universal' so if the unit uses 100uF at 16V and I can now get the same size at
25V or 35V I'll buy the higher Voltage to reduce inventory.

The ripple current directly relates to internal heating. NONE of the equipment ever gets to 105° C but the cap
can get hot from the charging/discharging currents. That's why caps bulge. The lower ESR (higher ripple
current) reduces that heating so it runs longer before it fails again. This is extra important in switching
power supplies. The linear supply in the Hafler would be OK with almost any cap of the right capacitance
and Voltage. The higher ripple current will likely cause a higher inrush current when turned on. The transformer
and diode bridge will be stressed a bit more for a few milliseconds but I promise you won't have a problem with it.
With a little expectation bias you can convince yourself it now sounds better. If the caps were bad enough it
may very well sound better.


That explanation was very much appreciated.
 

Talon

Member
2014-08-01 11:54 pm
Before I do anything with the caps I have to deal with the 50KAX2 ALPS volume pot:

alps-rk27-stereo-potentiometer-notched-shaft-high-quality-50-kohm.jpg


Two screws on the back so there might be a chance I'd be able to get this thing apart to clean it. If I have to replace it, there's the usual obstacles. It appears to be out of production as a 50k part, so I'd have to either switch to 100k or go look for NOS or at least a verified pull. I'm hoping to find NOS.

I took the list a bit further, organized it by type and ascending value within the type.

1.5 nF, polypropylene, 50v (2)
10 nF, polypropylene, 50v (4)
68 pF, polypropylene, 50v
120 pF, polypropylene, 50v
220 pF, polypropylene. 50v
470 nF, polypropylene, 50v


6.8 nF, 1%. polypropylene, 50v
25 nF 1%, polypropylene, 50v

15 nF, polycarbonate, 50v
47 nF, polycarbonate, 50v (2)
100 nF, polycarbonate. 50v (3)
470 nF, polycarbonate, 50v

22 mfd, electrolytic, 50v (3)
1000 mfd, electrolytic, 50v (2)

22 mfd, NP electrolytic, 50v (3)
4.7 mfd, NP electrolytic, 25v (3)
47 mfd, NP electrolytic. 16v
220 mfd, NP electrolytic, 10v (2)
1000 mfd, NP electrolytic, 6v

470 mfd, low ESR electrolytic, 25v (4)

I am in no way confident of either the Hafler parts list or my interpretation of all those typos. Couple that with the fact that Hafler like most vintage companies performed updates and alterations to a given design based on their own research plus customer feedback, and it looks like I'll be going over the chassis with a magnifier to verify everything.