Hafler DH-200/220 Mods

[tvrgeek]

No Mk II's out there right now. Missed last week I guess.

Pablo, what about your mods on the 200? There is not a lot to work with keeping the same topology. Lower noise parts, a bit of moving grounds off the board and back to the star point, a bit of PS work. Cascode the IPS maybe? Switch to JFETS on the input?

[tvrgeek]

Room for a third pair of transistors I guess. Switch to twin die you could get even more. Maybe use newer PS caps and split them into a pi filter. Just thinking here. Decent binding posts and input jacks. Mute and basic protection to replace the speaker fuse with a relay?

[tvrgeek]

Checked schematics, 535 Mk II has third output. Went back to standard bias spreader. No protection and no relays. The RB 951 has an advantage there. Mk I was DC coupled, Mk II, AC coupled, DC servo. Mk II is odd in that it takes the input off the inner of the LTP. Different amps that's for sure.

[pidigi]

sorry guys for the delay, I happen to work (way too much, lately) so the time left for the rest is really scarce...

You guessed almost all the mods "tvrgeek" (I suppose this is not your real name, but that's ok, neither Pablo is mine). The circuit remained the same. The pcbs are new, I used (matched) 2SK1058/2SJ162 as output devices, matched bjt also in the rest of the circuit (2N5551 here too), mf resistors, wima caps. The ground connections are completely different of course. Bigger trafo, sprague electrolytics, etc, etc... Basically all "standard" mods, after all. No magic here... Ah, forgot to say that the chassis is not the original one.

Yes, it is very difficult to "trust" the results of listening tests, even more if they are reported by someone else. That's way I was really surprised by your firm statements about the superiority of 150$ amps respect to Haflers. I'm not going to argue about the gift of your wife, but I cannot consider this gift as my way to judge power amps

So, as usual, that's my experience. And of course YMMV.

Ciao

Paolo (not Pablo)

[tvrgeek]

Apologies for miss spelling. Don't know used prices on the Continent, but I don't think I went over $150 for any of my Hafler's either.

Careful matching is a really big plus combined with 1% resistors. I bet the second harmonics dropped a bunch as it does not look like they matched the inputs originally, or at least not as close as you can get out of a batch of 20. Actually 40 for the 5551's and 5401's. The first batches were On-Semi, and I would consider them only semi-usable. Gain was "maybe". Fairchilds for the second batch. I think you have shown execution is as important as topology.

The Rotels of the vintage I speak are also simple on paper, but they just work better. I don't know why, which is what got me started in amps. There must be a reason. They don't use exotic parts either. Naim, MusicFidelity, and Creek are all highly respected and have very simple topology.

I have been scoping out newer transistors. 2SC970 and the like. Far lower noise, more linear. I am not going to spring for a new transformer, as lets face it, the transformer and heat sinks are half the cost. I would be more inclined to derate a 220 into a 75W by regulating the IPS and VAS rails at a lower level.

You are still left with no remote power control, no mute, or output protection system that one finds in newer amps.

It would be fun to know which gives more payoff, twin rectifiers and caps, or single rectifier and pi filter the same 4 caps. Did you do everything at once? I wonder, was the power supply and grounding what made the difference, or the exotic caps? Newer outputs?

[gootee]

One thing I also love is the Adcom GTP-500 II preamp/tuner that I also got from ebay. The tuner section doesn't work but the remote-controlled motorized volume knob does!

The remote also works to switch between sources.

[gootee]

Originally Posted by tvrgeek

I think he did the entire line. The larger have his spreader circuit from the schematics I saw. I find looking at cost constrained main-line examples to be interesting as you can see where their real priorities are vs. icing on the cake. 1K into One? No wonder the sub guys love them. Not scrimping on outputs I suspect is one of those priorities I mentioned.

Off to look for a Mk II.

Yup. 1KW into 1 Ohm, continuously, for each of two channels. It's right on the specifications page in the official service manual.

The seller's last line in the ebay listing was "She'll bring the thunder!".

I haven't tried one-Ohm loads, yet, though. My MG-12s are rated at 4 Ohms and the 585 is rated for 400 Watts into 4 Ohms. But I'm sure that I use only a small fraction of that.

I'm thinking that it might be very interesting to try paralleling four identical pairs of Magnepan 4-Ohm speakers. They seem to be at their clearest at lower volume levels, which is not extremely low (more like the "natural-sounding" level) but the MG-12s at least are not quite as good-sounding, I think, when they are pushed to extremely-high SPLs. So maybe with more than one pair, I could get much higher SPLs but with the total clarity of the lower volume level for each one.

Or imagine paralleling eight pairs of 2Ces, with one amp! <grin>

I wonder if those two scenarios would just be a mess, or if there would be a way to get a good soundstage image, if they were all in one room and devoted to one stereo listening position. I could see experimenting with stacking the MG-12s vertically in pairs, too!

Maybe then I'd want another 585 or two, so I could biamp or triamp! <grin>

Actually, I did purchase two of the 545 II amps, and the 535 II and the DH-220, thinking I might biamp or triamp some things. I even have schematics and everything I need (probably x100-to-1000) to build an analog active line-level crossover that would duplicate the (two-way) MG-12s' original crossover characteristics. I guess I need to spend less time at the computer and throw those things together and try it!

But for now, back to some Steely Dan!

[pidigi]

Originally Posted by tvrgeek

Apologies for miss spelling.

hey, c'mon, no need to apologize!

Quite often I find myself doing a "cut and paste" when replying to an email in order to not spell wrongly the names

Don't know used prices on the Continent, but I don't think I went over $150 for any of my Hafler's either.

Actually the number of Haflers you can find used around here is not high. That's why I decided to go for a clone instead of modifying one

Careful matching is a really big plus combined with 1% resistors. I bet the second harmonics dropped a bunch as it does not look like they matched the inputs originally, or at least not as close as you can get out of a batch of 20. Actually 40 for the 5551's and 5401's. The first batches were On-Semi, and I would consider them only semi-usable. Gain was "maybe". Fairchilds for the second batch. I think you have shown execution is as important as topology.

I agree with you. The first clone I have built wasn't using matched 5551/5401 nor 2SK1058/2SJ162. The result was not so impressive. The rest of the amp (grounding, caps, etc) remained the same in the second clone, but using matched devices really gave a big step forward

The Rotels of the vintage I speak are also simple on paper, but they just work better. I don't know why, which is what got me started in amps. There must be a reason. They don't use exotic parts either. Naim, MusicFidelity, and Creek are all highly respected and have very simple topology.

And this is really fascinating, quite often a simple topology gives best results. Simple, but not naive, and I think that this is the secret!

I have been scoping out newer transistors. 2SC970 and the like. Far lower noise, more linear. I am not going to spring for a new transformer, as lets face it, the transformer and heat sinks are half the cost. I would be more inclined to derate a 220 into a 75W by regulating the IPS and VAS rails at a lower level.

Nice idea the 2SC970, did you manage to try them?
I was also thinking about a low power version, mainly to test the difference in using a pair (or three as I do in the latest clones) output devices

You are still left with no remote power control, no mute, or output protection system that one finds in newer amps.

yeh, protection IS the main weak point of this amp. But it is quite easy to install a simple circuit to perform mute/output dc protection without spending too much/degrading the sound (not like the original fuse....).
In my experience a preamplfier is needed in order to get the best from them. So remote control is done in the preamplifier.

It would be fun to know which gives more payoff, twin rectifiers and caps, or single rectifier and pi filter the same 4 caps. Did you do everything at once? I wonder, was the power supply and grounding what made the difference, or the exotic caps? Newer outputs?

I can't say exactly, but my feeling is that grounding/layout are quite important. "Exotic" caps (well, wimas are not really exotic), 1% resistors, output devices are "second order" improvements. As I said, a big step forward is given from matched bjts.

There are still some other improvements that I would like to test (different bridge rectifier, changing VAS transistors, regulating the voltage of the board, etc). And then, the next step is trying to change something in the topology.

The "perfect" amplifier doesn't exist. When you reach a good quality, then quite often it becames a matter of individual taste. A friend of mine cannot live without extreme micro details but he doesn't care too much for a correct soundstage, for example.
For this reason there are on the market hundreds of different power amps..

Ciao!



Paolo

[tvrgeek]

I have not tried the new parts. I am working my way through analysis trying to actually understand what I am doing and why. Q7, 9, 10 and 11 in the 120 need more average power than the 907/1085 should provide so I am looking at other parts. KSA916/KSC12316 or 2SA695/2SC2235 maybe. I am waiting on more books to learn a bit more about the transistor parameters so I can pick them with more than just SOA and NF. I want to pick current production parts.

In my study, it turns out good protection circuits are really hard. Big second on the preamp as you must have a low Z source for lowest noise through the IPS. By remote, I meant remote power on and off.

[tvrgeek]

gootee,
Don't Maggies get really ugly when over driven and have a habit of going "
poof"? Maybe two pair of Peerless XLS's as woofers. They are good clear up to 200 Hz or so which would be a big help to the MG-12's. ( old paper ones, no idea about the new metal cones. I don't like metal cones). I am sure the guys in the sub threads could come up with ideas to tax the Adcom while getting every cop in town knocking on the door which you won't be able to hear anyway.