Hafler DH-200/220 Mods - Page 52 - diyAudio
Go Back   Home > Forums > Amplifiers > Solid State

Solid State Talk all about solid state amplification.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 16th November 2007, 10:08 PM   #511
diyAudio Member
 
Stormrider's Avatar
 
Join Date: Apr 2006
Location: Connecticut
Blog Entries: 2
Send a message via AIM to Stormrider Send a message via MSN to Stormrider
They may be faster, but as anatech is saying they are not better. The Multi-turn pots, as i understand it, do not handle the current levels very well, and the contacts age faster. So, what happens when a bias pot goes short circuit? Bad things that can be avoided.
__________________
AJ
  Reply With Quote
Old 16th November 2007, 11:09 PM   #512
djk is offline djk
diyAudio Member
 
djk's Avatar
 
Join Date: Feb 2001
Location: USA
It's when it goes open that I worry about.
__________________
Candidates for the Darwin Award should not read this author.
  Reply With Quote
Old 16th November 2007, 11:33 PM   #513
diyAudio Member
 
Join Date: Feb 2001
Location: Columbia, SC
I am always amused by claims that multi-turn pots are somehow inferior, usually accompanied by dark hints that they will come unstuck and roast your amp.
I've been in the mainframe end of the computer industry for over thirty years. Mainframe computers make a fetish out of reliability to an extent inconceivable to people in the audio industry, or indeed to anyone in any industry short of health care. In today's jargon, they speak of multiples of nines of up-time. That translates as 99.999% or better. Don't speak to me of networks. Networks are junk. Toys. Nobody in their right mind uses networks for "mission critical" work. Repeat after me: Mainframes are reliable. Period. They have to be. Period.
For all the time that I've been around mainframes, I've been collecting bits of cast-off computer hardware so as to salvage various parts. Transformers, caps, an occasional heatsink, that sort of thing. Having disassembled dozens of these power supplies, I've seen first hand what IBM, Amdahl, Digital Equipment, Storage Technology, et. al. put into their regulators and so forth.
So what's in them?
You guessed it...
Ten and twenty-turn pots.
I think I remember seeing one...count 'em, one single turn pot. And it was in something non-critical like a regulator for the indicator lights.
So, please, desist with the mutterings about reliability. If anything, comparing the reliability of audio equipment with the reliability of mainframes, a strong case could be made for replacing all single-turn pots in audio gear with multi-turn pots.
As with anything, you can abuse a part. It's no challenge to pop a 30A MOSFET, for instance, in spite of the fact that the 30A rating sounds quite impressive. Similarly, you can fry a pot, too. But is it the pot's fault if someone designs poorly?
Poor logic, that.

Grey
  Reply With Quote
Old 17th November 2007, 01:13 AM   #514
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
Hi Grey,
Multi turn controls are less reliable than their 270 counterparts. It is that simple. There is not enough contact pressure and they oxidize like everything else on this planet of ours.

I have worked on those computer power supplies. The most reliable control is that metal thing with five solder joints and a wire wound element. The second most reliable are the normal single turn jobs we all see, the larger ones.

There is no reason for a multiturn control in a computer power supply. There is in an A/D or D/A or any other above average fine adjustment. The T&M industry uses more controls than either of us have seen. Well, that's not entirely true as I did work in a calibration facility. I replaced TONS of those darn things. Never a single turn unless it was physically destroyed.

The T&M industry has gone to uP adjustable constants in order to eliminate multiturn controls (they do not hold cal very well). Ever hear of a "Trim Dac"?

So, unless there is a particular reason why you need extremely fine adjustments (bias current ain't one of them), a single turn is king as long as you've used your brain. They are easier to use too, you can get a tool into them.

The main difference between the two types are contact pressure. It should be blazingly obvious to all which one is more reliable (you know, the one that resists going open). In some multiturn controls, the wiper moves along a very small element where the wiper movement is geared way down. These are the worst of the multiturn controls. In other words a 5 turn is better than a 10 turn is better than a 25 turn - all other things being equal.

-Chris
__________________
"Just because you can, doesn't mean you should" my Wife
  Reply With Quote
Old 17th November 2007, 01:18 AM   #515
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
Hi Grey,
A question just occurred to me.
Quote:
Don't speak to me of networks. Networks are junk. Toys. Nobody in their right mind uses networks for "mission critical" work. Repeat after me: Mainframes are reliable. Period. They have to be. Period.
Why are you talking about mainframe reliability? Network reliability?

BTW, Telecom switches are very reliable. More so than your average computer. Mind you, they are computers too. Like a main frame only smaller. Not that Nortel junk, real Telecom switches. The redundant ones switching OC circuits.

Okay, we are talking on the same scale. So we can drop the computer references.

Now, Cal lab standards and calibrators need to be reliable. That would be the Test and Measurement Industry (T&M).

-Chris
__________________
"Just because you can, doesn't mean you should" my Wife
  Reply With Quote
Old 17th November 2007, 02:12 AM   #516
diyAudio Member
 
Join Date: Feb 2001
Location: Columbia, SC
I bring up mainframe reliability because it's based on adjustments made with multi-turn pots. I thought I'd made that obvious...perhaps not.
If you don't like them, fine, don't use them, but mainframe computers serve as a ready refutation of your argument that multi-turn pots are unreliable. After all, people who spend millions on a single chassis are unlikely to be particularly pleased if it dies because IBM used a cheap pot; the computer industry was using mil-spec parts long before audio manufacturers ever heard of the term.

Grey
  Reply With Quote
Old 17th November 2007, 02:42 AM   #517
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
Hi Grey,
No, I understood what you said clearly. Some of your post had no point though.

Quote:
I bring up mainframe reliability because it's based on adjustments made with multi-turn pots.
So, what was I looking at??? The ones I saw were single turn. Proper circuit design would eliminate the need for a multiturn control. If they are correcting for voltage drop across the cable, they are poor designers. Ever hear of remote sensing? You've made no point here Grey, not unless you can tell us specifically what those adjustments were for.
Quote:
but mainframe computers serve as a ready refutation of your argument that multi-turn pots are unreliable.
Not even close! Working in a calibraton lab (where your meter is sent to be certified) trumps that end use adjustment. The stuff I was working for was probably 4:1 or better than your limits. (Standard test uncertainty ratio). We found that equipment that depended on multi turn controls did not hold their calibration well over a year period. Most really needed a 6 month cal cycle. As an example, would you agree that a Fluke 87 was an appropriate instrument to use to set up those adjustments? Single turn controls inside that guy for calibration. How about an HP 34401A? Trim Dacs for closed case calibration and it's ability to hold calibration.

Face it. The one thing a calibration technician does not want to see are mutiturn controls. They allow for fine adjustment over a wide range. You do not need that range in your adjustment for bias current.
Quote:
the computer industry was using mil-spec parts long before audio manufacturers ever heard of the term.
No, computers went digital to be reliable. Analog computers used far more expensive parts. Face it. Any manufacturing facility does what it can to reduce costs. Besides, who ever said single turn controls were cheap?? They aren't, and they are very high quality. I don't think you are using facts to argue your case Grey. You have been mistaken on you viewpoint here. I think it's more a "high end tweaker" bias you are displaying.

By the way, "mil-spec parts" parts don't sound good. They are designed to be reliable under extreme circumstances. There are "mil-spec" single turn controls also. Same for resistors (carbon comp!! ) Are you telling me that "mil-spec" carbon comp resistors, ceramic caps, tantalum caps and similar things are on the audiopile tp 10? 'Cause they ain't. Our audio equipment (or your beloved main frames) are not expected to withstand extreme G forces, radiation or temperature extremes. The parts IBM and others are industrial spec parts. There is zero benefit for them to use the full temperature range parts for a computer that is housed in it's own climate controlled room.

What makes your main frame computer reliable is not the multiturn controls. It's the very conservative design and the environment spec.

Not trying to pick a fight here, but I refuse to allow you to mislead our members and propagate a myth. Believe what you like, but you might want to do some research first.

-Chris
__________________
"Just because you can, doesn't mean you should" my Wife
  Reply With Quote
Old 17th November 2007, 03:36 AM   #518
diyAudio Member
 
Join Date: Feb 2001
Location: Columbia, SC
Chris,
You and I appear to live in very different worlds. I don't know about the one you're in, but in the one I live in, the first electronic computers were built for the US military; used for calculating ballistic tables for artillery, then, later, for code cracking (pre CIA, so this would have been OSA). Yes, they were analog at first, but digital came along as soon as it was practical. What persisted was a relation between the military and computers. The military has always been one of the biggest users of computation and mil-spec parts were part of the game if you wanted to sell to them. Why is that so difficult to understand and why are you getting so wound up?
IBM did not have two product lines, one for the military and one for commercial use. They sold the same mainframes to the military that they sold to large corporations, universities, etc. Now, every once in a while the military would special order a piece and IBM would create a one-off research project that would create a trickle-down effect in their regular product, but then anyone with enough money could commission a piece...it's just that the military had deeper pockets and a burning desire to model thermonuclear detonations, so they had more of the one-offs (these days, most if not all of the code cracking is handled by the CIA and NSA). But all the mainframes were built to the same standards.
(In fact, we've got some computers here that were originally at NASA, later donated to us. Some of them still have the NASA inventory stickers on them. Lots of the stuff we get is donated, although our most recent mainframe acquisition was purchased new, direct from IBM. Just came online a few weeks ago.)
The first mainframe installation I worked at was housed in the basement of the physics building at the University of North Carolina. An unremarkable location until you realized that it was radiation-hardened. The computer had been bought by the US government for the census, then donated to UNC when they were done with it. So not all government purchases were military, although they thought highly enough of this particular installation to build it in a fallout shelter, just in case. Personally, I always figured that if we had a nuclear war, any previous census results would be null and void, what with all the charred bodies, etc. But maybe that was the point...maybe the plan was to do a census as soon as the bombs and missiles quit falling. Who knows?
(Still had the signs next to the doors when I worked there. Sobering. Particularly since we were in the thick of the cold war in those days.)
No, I haven't gutted the new mainframe to see what kind of parts they're using these days. I think they'd notice...
Where on earth you got the "sounds better" part, I don't understand. I never said such a thing. I doubt seriously that MIT polystyrene caps are mil-spec, but they sound better than more "normal" parts, so I use them in tube gear. I certainly don't demand that parts be mil-spec and can't understand how you've worked yourself into such a lather over things I never said.
My one and only point is that multi-turn pots are not necessarily unreliable. I have not said that they sound better (or worse), nor have I said that they are more expensive (or cheaper), nor have I made any other claims other than their being suitably reliable.
If IBM feels that multi-turn pots are sufficiently reliable to hold power supplies to 5V (or 45V or 12V or 30V or 3V...depending on whether the power supply was for logic, tape drive motors, lights, etc.) then I daresay that they're probably going to do okay in a home-brew bias circuit.

Grey

P.S.: Why would Chapel Hill, NC be worth a fallout shelter for a US government mainframe installation? The story I heard was that a major telephone junction for the entire Southeastern US was situated just south of town. Whether it was true or not then (early '70s), I have no idea, but I doubt that they would maintain such a critical thing in one centralized location these days.
I always assumed that the siting of the computer was due to exactly that communication link--they'd have faster, more reliable access (think 300 Baud or so in those days) to incoming data from the surrounding region, but that's just supposition on my part.
  Reply With Quote
Old 17th November 2007, 04:30 AM   #519
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
Hi Grey,
Our experience certainly differs! I'm talking about things I have had my fingers into up to the present day. Our different worlds may be the same ones separated by time, who knows.

I can say with certainty that two things are true.

1.) Multi turn controls available to the average user (and worse, NOS!) are not as reliable as a single turn control. Contact area and pressure have a lot to do with this. I am sure you can find some mil spec parts that do not suffer from this as much, but they will not be used by the average person.

2.) Multi turn controls are not required to keep a power supply set to a voltage that can not be accomplished with a single turn control. Most industrial, high reliability power supplies, use single turn controls. Even the telecom systems the US Navy uses on their carriers have power supplies using single turn controls. They just use two supplies. I have one here that I'm looking at right now. Same ones used by your Homeland Security offices too (different model).

It is not very difficult to set your trim range so that the adjustment is not sensitive and it covers the entire required range. That is all I have said. I added that if "you" don't figure it out, "you" are being lazy. Too much adjustment range is a very bad thing.
Quote:
Where on earth you got the "sounds better" part, I don't understand. I never said such a thing.
I said that because we are talking about audio gear. When someone starts talking about using mil spec parts in audio equipment, it's because they generally attach some magical quality to them so they are supposed to sound better.

At any rate, I have seen ample proof where multi turn controls have generally been less reliable than the single turn variety. I've had to repair countless amps where a "modder" replace normal controls with multiturns only to have them fail. And finally, the failure rate in test equipment has been very high over time. The one example I pointed out to you is that techs are setting those supplies with voltmeters that no longer use multi turn controls due to reliability issues. I don't know what else I can say beyond that. I will say that your assertion that these parts are suitable for use in audio is misguided. The fact that you suggested that single turn controls were cheap hints to me that you are not immune to the popular misconceptions.
Quote:
After all, people who spend millions on a single chassis are unlikely to be particularly pleased if it dies because IBM used a cheap pot; the computer industry was using mil-spec parts long before audio manufacturers ever heard of the term.
This statement alone prompted me to think you were in the "mil spec is better for everything" camp. Glassy eyed and drooling. Remember that in the beginning, everyone used the same parts. Electronic computers did not exist yet, but audio devices did. Telephone gear did, and that is what really drove the audio market.
So please remember that pretty much everything else existed as an industry before computers became an industry (or even existed).

Finally, a power supply that requires a multi turn control to set the output voltage is very poorly engineered. You can see that often a couple precision resistors are all that is required with some of the newer regulator control chips. Even switchers. So a multiturn in a power supply tells me they are using really lousy voltage references.

I don't know Grey, you really have to peer into some newer stuff.

So, the reason I am digging in here is that you are wrong about multiturn controls. I have had to repair these failures time and time again. There is a cult status these parts have achieved also that you are only feeding. I can not allow you to misled people in this way. As you know, this is absolutely not personal. It has only to do with many, many years of observed failures.

-Chris
__________________
"Just because you can, doesn't mean you should" my Wife
  Reply With Quote
Old 17th November 2007, 01:04 PM   #520
diyAudio Member
 
jackinnj's Avatar
 
Join Date: Apr 2002
Location: Llanddewi Brefi, NJ
Can we get back to discussing Hafler.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Hafler dh110 mods anyone? amp_guy Solid State 5 30th January 2011 12:57 AM
Hafler xl-280 mods Audionirvana Solid State 5 11th November 2010 04:25 AM
Hafler DH 101 preamp mods Zen Mod Solid State 6 3rd May 2004 12:24 AM


New To Site? Need Help?

All times are GMT. The time now is 07:30 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2