Hafler DH-220, Bias and DC offset - need practical experience help

pkquat

Member
2016-04-13 4:06 am
I have an all original Hafler DH-220. Its in great shape, and was dormant for 10+ years. Both channels are fine, and run the same temp, about 108F ± a degree. The DC offset had crept up from 15-20 mV to near 30mV over the past year.

I know about basic electronics and some troubleshooting, and usually have no issues with good instructions, but I don't know much about circuits or design etc.

Checking / adjusting the bias and DC offset are well described in the manual, but some additional practical experience would be helpful. I am turning the upper pot to adjust the DC offset. I only have one MM so I check one channel at a time.

When I turn the amp on, after the initial surge the bias starts around 275mA. The bias always creeps up to around 285mA after 15-20 seconds and to around 295 after a minute. It was still climbing, but very slowly. I am not sure if this is normal or not. If it is, how long do you leave the amp on? How long until it settles? After turning the amp off and letting the values drop to around 0.2mA, if I power it on again, it starts around 285mA. If I disconnect the leads after it drops to around 0.2mA and put the fuse in place briefly, when I reconnect things and power it up, it starts at 275mA.

How long does it take for the offset to settle? Are both channel offsets "connected" in their offset since it is a common ground amp? Do I need two MM's? It seems like once I adjust one side, the other changes. If I leave the amp on, and come back in 20-30min the values have also changed.
 
Adjust P2 until the desired quiescent current reading is achieved. Q9 should be thermally connected to the heat sink to keep good current tracking but the circuit is awful. (If P2 was between the base and emitter of Q9, there is less chance of an overheating problem) but it is not and leave it alone. When the quiescent is sorted out then achieve the DC off set. After that, check the amplifier with your speakers once you know there are no problems.
 

pkquat

Member
2016-04-13 4:06 am
That you for the reply. I understand that part. I am interested to know if the bias on the Halfers creep. If it does, will it eventually settle? Or is somewhere near an initial reading what you go with? Does the amp need to be at full operating time to adjust the bias? Might something be wrong with the amp, or multimeter causing the creep?

I know the DC offset is adjusted with the amp at temp. I am fighting the adjustments. It seems like the channels are connected somehow. Adjust the left channel and wait 2-5 min and it is still the same value. I will adjust the right channel. When I go back and check the left, it is has changed. It also seams is if I only adjust 1 channel, and wait 2-5 min the reading is nearly the same, within 5 mV bouncing. If I wait 30 min the nominal value can be different by 10 mV. From then on it is constant. If I turn it off let it cool, and then turn it on and get warm, the values are the very close to the same.
 
I don't think there's anything wrong with your amp or multimeter. Your "creep" is a (lack of) thermal tracking issue hinted at by Jon.
I've had three DH-220 amps and they all exhibit similar behavior to what you've noticed.

Remember, you have a common (unregulated) power supply between the two channels. It wouldn't be unexpected for your adjustment on one channel to cause a minor change on the other.

Cheers,

Dave.
 
The types of mosfets used in the DH220 have good inherent stability at that bias level. As long as it isn't climbing out of control (don't know if it's even possible with that amp), I'd not worry about it.

Also 30mV of offset, while not ideal, isn't really much of a problem. Into 4 ohms, that's only 7.5mA, not enough to bother a decent woofer and from an amp that can do about 1000x that much.
 
Bias will increase with time and then stabilize after around 45mins to an hour and the DC offset will change. Assuming clear air circulation around the amp.

I have the DH220 and the 9505. First connect the MM to one side to adjust the bias and put the lid in place as best as you can. Let the amp get hot for an hour until you see that the bias is not changing and then adjust. Do the same for the other side (since you have only one MM). Then adjust DC offset. Factory specs call for 275ma bias but the amp will sound far better at higher bias levels. You may adjust until heat sinks do not exceed 125F.

Nash
 

pkquat

Member
2016-04-13 4:06 am
WOW! Thanks guys. This is the information I was looking for. Thanks to Jon as well, but the terminology was beyond my knowledge.

I think my bias is good. Both sides are within 1-2°F I got both sides down to 0-5mV. I wish I knew both sides were connected sooner, and that while it settles quickly after the initial bump you have to wait much longer for both sides to totally settle. I pretty much figured that out, but what a :headbash: to get them dialed in. Thank you for the confirmation as well. 2 MM's would help, and a chart of how much one side effects the other would be even better.

I'll do a follow up post with a little story and some additional questions regarding bias checking and adjustment, but first I want to understand the process and the creep better.

For the bias, should I get the amp warm, turn it off, wait for 2 min (or more or less?) and then remove the fuse to check the bias, connect the MM leads, and then wait another XX min and make sure it is stabilized and then adjust it? Or take the fuse out when the amp is cold, and let it warm for 45 or so minutes? I didn't think it would get warm while the MM was in place. I guess I was mistaken?

I think 10+ years of the amp sitting dormant helped. I also never left it on for extended periods of time. I will probably see about checking some of the electrolytic caps (if that is possible) and replacing them. I'll have to find some thread on what all to check and what common replacements are. The transformer is starting to barely hum again. IDK if I can tighten it up any more.

Some people love the amp, and others pan it. The amp has impressed me. It keeps pretty close pace with a 150W @ 8Ω Krell 300i. Its driven some impressive loads and speakers so far.
 

pkquat

Member
2016-04-13 4:06 am
The longer story on the bias is that the first time I went to check it, I only had one clip, and held the probe on the other end. My only intention was to check it at that time. The left side went fine, but on the right side I broke contact, and re-established it. It says in the manual not to have intermittent contact. What issues or damage can it cause if you do? If I recall once the circuit is broken you can't really re-establish it without powering the unit off and the re-connecting things. Also when this happened there was a little ozone smell, and a touch of the burnt electronic smell. No smoke however, and the burnt electronics smell was not real strong or the complete burnt plastics.

When I powered it back up the right side was reading 400+mA. I checked the left and it was still around 275 at start up. The right stayed at 400+mA even after letting it sit. I turned it on and left it on for a little while, and the right side was running way hotter. Because of the creep I compared the sides cold and tried to match the right side intial start up values to the left sides. I moved the pot a significant amount but eventually dialed it back to the original position the best I can tell. It took a number of iterations of letting things power down, and then staring back up to match an initial reading.

Everything appears ok now, but is there anything I should check to see if something may have been damaged, or a resistor briefly overheated possibly changing the value?
 
I have 3 DH-220's that I've been using since the mid-1980's. I built two of them from kits. Since you stated that this amp hadn't been used for 10 years, I'd be suspicious of the state of any electrolytic caps. You might wan to replace them. Any 30 yr old electrolytic that sat dormant for 10 years is questionable. A Rotel amp I bought had about 60mV of output offset and there was no adjustment... I actually built the offset circuit that a DH-220 has, and installed it into the Rotel. 60mV is too high IMO. Now it's less than 10mV. Voltages often drift for the first few minutes the amp is turned on. I'd check and adjust it after 10 minutes of warm up.
 

pkquat

Member
2016-04-13 4:06 am
Any 30 yr old electrolytic that sat dormant for 10 years is questionable.

Whats the difference between a 30 yr old electrolytic that has been in use or sat dorman? I thought they suffered the same from aging, but often could dry out faster in hotter environment. I have two pieces of vintage gear that are still original and working well. With everything I hear about older caps I am surprised.

I was thinking about replacing the electrolytics at some point, but so far everything appears to be working fine, so I have not been in a rush. From what I have gathered they should be replaced with electorlytics although their are other types of caps that age better. Are their some that could be upgraded to a better cap?
 
Whats the difference between a 30 yr old electrolytic that has been in use or sat dorman?
Good question. Being in regular use does make a difference though, as it allows suspicious spots in the dielectric layer to reform. No use for some time may mean that the dielectric degrades enough for those suspicious spots to break down and be unable to fully recover, leaving the cap leaky. IOW, regular use may mask a cap that technically might be considered dead already. That's my take on it anyway.

Recently I encountered an 8-year-old PC that would always turn off again while booting Windows after not having been used for ~4-5 months, with no prior symptoms of anything being wrong. Thought it would be the PSU, nope, turns out it's the board. (Thankfully there was another spare PC of same type around, so a HDD swap later I was back in business. Resetting the CMOS password nobody knows required ripping everything apart twice though, how annoying.) It would barely get to the GUI when warmed up for a bit, not even that when cold. Sure smells like bad electrolytic(s) to me. I've been meaning to get an ESR meter for some time anyway. An old G31 board isn't worth anything much but it would be an interesting learning experience, and I might get to try a Xeon mod as well.

Speaking of PC matters, flash memory with a lot of cycles on it will exhibit shorter and shorter retention time as well, so a well-used SSD would eventually exhibit similar effects.
 
P3000, P1000?

I have had nine Haflers (after several Dynakits) and they have been very serviceable. But today I get no response adjusting bias on a P1000. The day before, attempting bias adjust blew my DMM ammeter's 2A fuse. The expected measurement is 100mA. Power supply rails are normal +\-49vdc, and nothing I know of has fried except the fuse. Two P3000s gave similar experiences. Anyone have advice?
 

msb64

Member
2008-02-22 6:10 pm
I know the DC offset is adjusted with the amp at temp. I am fighting the adjustments. It seems like the channels are connected somehow. Adjust the left channel and wait 2-5 min and it is still the same value. I will adjust the right channel. When I go back and check the left, it is has changed. It also seams is if I only adjust 1 channel, and wait 2-5 min the reading is nearly the same, within 5 mV bouncing. If I wait 30 min the nominal value can be different by 10 mV. From then on it is constant. If I turn it off let it cool, and then turn it on and get warm, the values are the very close to the same.

I experienced the exact same thing working on various DH200 and DH220 models.
I would let the amps run for at least 45 minutes or so to get to full temperature and stabilize.
My solution to reduce the time for the tweaking back and forth for the 2 channels was to use 2 separate good/verified accurate meters, each with clips on all probes.
That way I could safely run them for longer durations and maintain good contact.\

To allow for fine tuning the dc offset on my DH200s, I just duplicated the circuit from the DH220s and added it to each side on a small perf board (see photo).
The coolest part about this DH200 was that it was a wreck when I got it, it was so bad that I wouldn't even plug it in when I got it.
It was poorly constructed, messy wiring, terrible solder joints, beat up and rusted chassis, just an all around disaster.
I ended up going through every piece by either replacing old components such as caps and resistors, cleaning components if original, new insulators and thermal compound etc.
Also did some wiring, ground and cord changes, and used good cost effective parts from places like Mouser & Parts Express.

As for the bias settings, older upgrade mods used to push them up to as much as 425mA.
I like to be on the conservative side, so keep mine in the 275 - 300mA range.
Check out the DH 200/220 Mods thread, takes awhile to go through it all but there is some really good info that you can apply to your amp.
 

Attachments

  • DH200 Rebuild.jpg
    DH200 Rebuild.jpg
    282.4 KB · Views: 209

scott wurcer

Disabled Account
2004-01-26 3:03 pm
Belmont MA
As for the bias settings, older upgrade mods used to push them up to as much as 425mA.

I did many of Walt Jung's Pooge mods and ran them at 425mA (~50W idle per channel) with no problems. I don't listen at very high volume and never had a problem. Funny thing just last week I located the last lost DH220 that I built in the 80's in my sons basement where it has been under water several times due to flooding a fun restoration project.
 
Last edited:

msb64

Member
2008-02-22 6:10 pm
Pretty sure mine came from a flooded home too, somewhere down in the U.S.
I also did the Pooge mods, or at least attempted several variations of it, with TRT Wondercaps, metal film resistors, remote torroidal power supply, larger caps etc.
I only had photocopies of the (2nd?) article, but later on tracked down a new copy of the magazine.
Back then I had some decent gear to test with, including Infinity IRS Delta speakers which really helped to reveal what it was capable of.
Still running a pair of extensively modded mono DH200 Haflers as my mid/high amps today, along with a Threshold S350e for bass.
Lots of potential in those amplifiers for sure, I still have at least 3 more waiting for some TLC.
 

phase

Member
2004-10-04 11:59 pm
Not that many amps that will be viable candidates for restoration post flooding, but now wasn't David Hafler in the navy?
Raises your standards/expectations apparently working with tactical grade equipment.
Adjusting the bias by bypassing the fuse is a very potentially reckless way to get those numbers, but it works only with the clip on leads and lots of care.
 
Bias of a 220

I have a Haffler 220 that I'm currently working on and it seems to be working well at 100 ma of bias. While I haven't measured the output wattage on it or distortion yet my last one measured 140 watts at .001% Thd and is sounding great on a pair of Klipsch.

My question is,
At 100 ma of current instead of the traditional 275ma is their any risk of running the amp this way?
Thank you.