Yes... trimpots are evil.
The thermo-electric voltage from wiper to element is also a pain.
Some dirty pot tricks:
1) don't use one... calibrate in software.
2) use a matching style sacrificial pot right next to the real one in your circuit. arrange it in such a way that the wiper currents flow in opposite directions... this will help to cancel thermal EMF's.
3) don't use one.
4) In voltage dividers... never use a pot as a potentiometer... always configure it as a rheostat.
5) don't use one.
6) the tempco's are horrible... use proportionate values in each branch of a voltage divider to help cancel the effects.
7) don't use one.
8) the tempo's are still horrible... the wattage rating only implies that amount of power which will keep the smoke inside the device. Stay under 5% of the wattage rating where you can.
9) don't use one.
10) forget about multi-turn pot's. They are really just single-turn pots with a gearbox... so you think you are doing a great job. The gearbox is loaded with mismatched coefficients of thermal expansion... an inaccurate temp sensor at best.
11) don't use one.
12) use wire-wounds when you can... this avoids much of the horribleness.
13) don't use one.
14) always sweep the wiper 5 or 10 times from min to max... this knocks down the lumps and let's you think you did a better job.
15) don't use one.
16) always tap the pot when you're done adjusting it to relieve stiction. This will kill your setting though.
16) Did I remember to say, "don't use one"?
17) Vishay metal foils are excellent... try finding one... and then try to buy it.
The thermo-electric voltage from wiper to element is also a pain.
Some dirty pot tricks:
1) don't use one... calibrate in software.
2) use a matching style sacrificial pot right next to the real one in your circuit. arrange it in such a way that the wiper currents flow in opposite directions... this will help to cancel thermal EMF's.
3) don't use one.
4) In voltage dividers... never use a pot as a potentiometer... always configure it as a rheostat.
5) don't use one.
6) the tempco's are horrible... use proportionate values in each branch of a voltage divider to help cancel the effects.
7) don't use one.
8) the tempo's are still horrible... the wattage rating only implies that amount of power which will keep the smoke inside the device. Stay under 5% of the wattage rating where you can.
9) don't use one.
10) forget about multi-turn pot's. They are really just single-turn pots with a gearbox... so you think you are doing a great job. The gearbox is loaded with mismatched coefficients of thermal expansion... an inaccurate temp sensor at best.
11) don't use one.
12) use wire-wounds when you can... this avoids much of the horribleness.
13) don't use one.
14) always sweep the wiper 5 or 10 times from min to max... this knocks down the lumps and let's you think you did a better job.
15) don't use one.
16) always tap the pot when you're done adjusting it to relieve stiction. This will kill your setting though.
16) Did I remember to say, "don't use one"?
17) Vishay metal foils are excellent... try finding one... and then try to buy it.
Thanks for the answers. I think I have asked once before with no results at all. I guess one has to ask when the right guys are online.
So metal film or tatntalum sounds to be the choice if available. For most of us that would label under the exotic though, since they seem not so common. I will keep in mind that they migt be worth looking for in critical situations, though. Any opinions on carbon vs. cermet? Those are usually the only types easily available.
SY has an excellent point. In contrast to the material in the film it is not so easy to find out from tech. specs. though. I guess the best one can hope for is that more expensive pots are on the average better designs in this respect.
Poobah, you had many very interesting points. Especially the odd numbered ones, although I think I knew most of those already.
Of course, as long as we are doing DIY and not designs for mass production, there is often also the option to first use a trimpot to find out the right value and then replace it with fixed resistors, assuming it is a parameter that need not be readjusted later. I guess that would be too expensive in production though, even for most of Johns designs.
So metal film or tatntalum sounds to be the choice if available. For most of us that would label under the exotic though, since they seem not so common. I will keep in mind that they migt be worth looking for in critical situations, though. Any opinions on carbon vs. cermet? Those are usually the only types easily available.
SY has an excellent point. In contrast to the material in the film it is not so easy to find out from tech. specs. though. I guess the best one can hope for is that more expensive pots are on the average better designs in this respect.
Poobah, you had many very interesting points. Especially the odd numbered ones, although I think I knew most of those already.
Of course, as long as we are doing DIY and not designs for mass production, there is often also the option to first use a trimpot to find out the right value and then replace it with fixed resistors, assuming it is a parameter that need not be readjusted later. I guess that would be too expensive in production though, even for most of Johns designs.
I'm surprised that poobah didn't mention the other thing with trimpots. Never pass all of your current through them. Always use the trimpot in series with a fixed resistor, all in parallel with a fixed resistor. That way, you can ensure that the trimmer only gives the exact range of adjustment you genuinely need and that all its awful problems are swamped by the majority of the current going through the fixed resistors. It's not usually difficult to get a 10:1 ratio of current, thus effectively dividing the trimpot's problems by ten, and making them useable.
I wrote a little spreadsheet that plots the variation in total resistance against wiper rotation for the above scheme so that I can plug various values in to see if I like the look of the law and the range of variation. It allows me to force a standard value of trimpot to give exactly the weird range of variation I need.
I wrote a little spreadsheet that plots the variation in total resistance against wiper rotation for the above scheme so that I can plug various values in to see if I like the look of the law and the range of variation. It allows me to force a standard value of trimpot to give exactly the weird range of variation I need.
Christer,
What I've done often where extreme accuracy is needed is to place 2 resistors in parallel.
Load the board in production with only one of the resistors about 1% (or whatever the math says) over the target value. Then for calibration, derive a chart for resistor values versus calibration value, and hand load the second resistor. This actually is not too bad of a method... doesn't take that long.
OH!
18) Don't put trimots on a board because some nitwit will have to screw with it.
I actually had moron call me once wanting to return a circuit board that he had burned up. His complaint was that there was a stripped screw "that wouldn't draw up tight"... the small screw on a muti-turn pot.
19) Always reverse the screw slightly after adjusting a muti-turn pot... this relieves residual tension in the mechanism and prevents the setting from changing after the first jolt or temperature excursion.
20) Plan for an open wiper (back to the rheostat)... configure your circuit such that the ejection seat won't fire in the event of an open.
What I've done often where extreme accuracy is needed is to place 2 resistors in parallel.
Load the board in production with only one of the resistors about 1% (or whatever the math says) over the target value. Then for calibration, derive a chart for resistor values versus calibration value, and hand load the second resistor. This actually is not too bad of a method... doesn't take that long.
OH!
18) Don't put trimots on a board because some nitwit will have to screw with it.
I actually had moron call me once wanting to return a circuit board that he had burned up. His complaint was that there was a stripped screw "that wouldn't draw up tight"... the small screw on a muti-turn pot.
19) Always reverse the screw slightly after adjusting a muti-turn pot... this relieves residual tension in the mechanism and prevents the setting from changing after the first jolt or temperature excursion.
20) Plan for an open wiper (back to the rheostat)... configure your circuit such that the ejection seat won't fire in the event of an open.
Chris... you got me laughing out loud!
Did you mess with that?
No!
Then why is blah blah blah?
I dunno...
That's one of those, "wish I had a dollar for..." things
EC,
Good points...
I keep a resistor chart on the wall. After the pot value is determined, I scale all the resistor values up or down to accomodate a standard value pot. But you can't always do that now can you?
In the same vein, I have found that if you size a pot so that ALL circuits will calibrate... you know, this biggest A, with smallest B, and the worst C etc..., you wind up with pots that are overly sensitive and occupy too much of the circuit; usually by about a factor of 3 or 4. Now, I love Monte Carlo and statistics as much as the next guy, but it's worth it sometimes.
20) Use a smaller pot and be prepared to rework 1 out of 10,000 circuits that won't cal because you ran out of pot.
Someone's gonna run with that...
Did you mess with that?
No!
Then why is blah blah blah?
I dunno...
That's one of those, "wish I had a dollar for..." things
EC,
Good points...
I keep a resistor chart on the wall. After the pot value is determined, I scale all the resistor values up or down to accomodate a standard value pot. But you can't always do that now can you?
In the same vein, I have found that if you size a pot so that ALL circuits will calibrate... you know, this biggest A, with smallest B, and the worst C etc..., you wind up with pots that are overly sensitive and occupy too much of the circuit; usually by about a factor of 3 or 4. Now, I love Monte Carlo and statistics as much as the next guy, but it's worth it sometimes.
20) Use a smaller pot and be prepared to rework 1 out of 10,000 circuits that won't cal because you ran out of pot.
Someone's gonna run with that...
anatech said:
Because it is usually the only thing that is easy to mess with. Look! Here is something I can change. Wonder what happens if I do?
Poobah, yes one will often need two fixed resistors in parallel to replace a trimpot, but you have a very good point in starting with one of the fixed resistors in place. That makes sense also for single DIY things, since it gives better precision when adjusting and measuring the trim pot.
Hm, I think you actually forgot one more point, which you all know, but since there are most certainly more people just reading the thread:
When connecting a (trim) pot as a rheostat, connect the "loose" end to the wiper so you won't get an open circuit if the wiper fails. This in combination with that EC8010 recommended might be difference between a circuit self-destructing or not in some cases.
BVT2000 "C" format time base corrector
Ah, Japanese tweaks! I met an awful lot of broadcast equipment that had tweaks with far too much range on them. I remember one where the chrominance (posh for colour) path had an adjustment range from 0 to 200%. Anything more than 5% adjustment about nominal in that area should have been cause for alarm bells. Naturally, this particular adjustment was hair-trigger and drifted (and was difficult to get at). Designed by "those terribly nice people at Sony..."
poobah said:
Ah, Japanese tweaks! I met an awful lot of broadcast equipment that had tweaks with far too much range on them. I remember one where the chrominance (posh for colour) path had an adjustment range from 0 to 200%. Anything more than 5% adjustment about nominal in that area should have been cause for alarm bells. Naturally, this particular adjustment was hair-trigger and drifted (and was difficult to get at). Designed by "those terribly nice people at Sony..."
Hi EC8010,
and DC offsets).
Of course, the Revox A-77 trimmers simply fell apart when adjusted (touched). No going back now! Yes poobah, you could say you just ran out of pot.
-Chris
Everything the Japanese make is adjustable. Precision costs far too much to design in. Japanese R-R machines - every darn thing is on a spring loaded screw. Studer and Revox, most times it fits and lines up. Slight shimming maybe. Electrically, always far too much range! Yamaha was perhaps the worst for excessive range (bias controls!!
and DC offsets).Of course, the Revox A-77 trimmers simply fell apart when adjusted (touched). No going back now! Yes poobah, you could say you just ran out of pot.
-Chris
poobah said:OFF TOPIC:
Hey Chris,
Is a ReVox B225 CD player worth fixin?
ya nutz?
ReVox?
http://www.revoxonline.ch/index-download.htm
http://www.revox.name
- Status
- Not open for further replies.