Great! Does that include the calibration of the actual pot values for the controllers binary position code?
Also I assume since you are connecting the pot as a two terminal resistor you are grounding the wiper, to minimize stray capacitance.
ONE NOTE for other folks. Never try to measure the minimum value of a pot with a VOM. A 50K 1/2W pot is designed for a maximum current of 3.16ma. The common method to measure resistance is to pass a current through the resistor under test and then measure the voltage. At the low ohms range on some meters this can be a current greater than the pot is designed to handle! So before you check stop resistance be sure your meter uses a test current that is safe!
Simon, I apologize if you want to go on and design the control electronics as well. I don't have to, and I will leave that to other experts at the disposal of the company that I design for. However, I still don't have an answer for the optimum series/parallel resistive switch array to use with a 1K-2K pot to extend it to 100K. Any ideas?
John,
Not wanting to be a pita, but if your control range extends up to 100k, why do you want to use a 1-2k pot? Really am curious.
jd
Last edited:
A few relevant notes:
1) I checked a few 10 turn precision pots in my collection. Much to my surprise the capacitive component was dominant even on the 100 Ohm sample. Its very small so they can be considered purely resistive up to 140 KHz.
2) the linearity is usually .1% or better so its easy to get multiple pots to track close enough for any real application. Many years ago I had an early DVM that consisted of a 10 turn pot, a nulling amp + servo and a mechanical readout. It held .1% accuracy adequate for a number of military contracts. It was fun to watch but slow.
3) a servo tied to a microprocessor will easily handle setting issues if the pots track (this is how its handled in the NuForce control)
John has a real issue re: switching load resistors. If 47K is an option then either the pot is switched in parallel or in series with the resistor. Then the switch (relay) becomes part of the circuit. The gold contact TO5 relays are very good, and the small size ensures small reactive effects. But they are very costly and are yet another contact in the chain. The good part is the contact resistance is in the 30 mOhm range normally. the truely obsessive would want to get them with copper leads. The sample I pulled from my collection had kovar (magnetic) leads.
1) I checked a few 10 turn precision pots in my collection. Much to my surprise the capacitive component was dominant even on the 100 Ohm sample. Its very small so they can be considered purely resistive up to 140 KHz.
2) the linearity is usually .1% or better so its easy to get multiple pots to track close enough for any real application. Many years ago I had an early DVM that consisted of a 10 turn pot, a nulling amp + servo and a mechanical readout. It held .1% accuracy adequate for a number of military contracts. It was fun to watch but slow.
3) a servo tied to a microprocessor will easily handle setting issues if the pots track (this is how its handled in the NuForce control)
John has a real issue re: switching load resistors. If 47K is an option then either the pot is switched in parallel or in series with the resistor. Then the switch (relay) becomes part of the circuit. The gold contact TO5 relays are very good, and the small size ensures small reactive effects. But they are very costly and are yet another contact in the chain. The good part is the contact resistance is in the 30 mOhm range normally. the truely obsessive would want to get them with copper leads. The sample I pulled from my collection had kovar (magnetic) leads.
The T0-5 relay is a given in this project. It is just using the MINIMUM number of contacts and fixed resistors to extend the range to 100K (perhaps a little rough, after 10K, with a 1K or 2K 10 turn precision hybrid wirewound pot as the fine adjustable reference for loading up to 1K or so. Demian, did you specify copper leads for your project?
PMA:
As you probably know there are many different relay contacts used for different applications. No simple answer. And some solutions that would be great for some applications will be unusable for others. For example silver is very unreliable for low energy contacts. It will sulfide at a predictable rate and the surface contamination needs a certain minimum current to breakdown. Gold or mercury work best at low levels. Mercury has the longest service life, gold the lowest contact resistance. Mercury is only available in reed relays and is considered hazardous so not really available.
My experience is that the lowest contact resistance for mercury is around 30 mOhms, gold in the 5 mOhm range but degrades faster with use. And measuring contact resistance in the milliOhm range is hard at low energy levels.
As you probably know there are many different relay contacts used for different applications. No simple answer. And some solutions that would be great for some applications will be unusable for others. For example silver is very unreliable for low energy contacts. It will sulfide at a predictable rate and the surface contamination needs a certain minimum current to breakdown. Gold or mercury work best at low levels. Mercury has the longest service life, gold the lowest contact resistance. Mercury is only available in reed relays and is considered hazardous so not really available.
My experience is that the lowest contact resistance for mercury is around 30 mOhms, gold in the 5 mOhm range but degrades faster with use. And measuring contact resistance in the milliOhm range is hard at low energy levels.
Is the requirement then for instance to vary the resistance at the nominal 47k to 47k +/- 1k? Is that the setability required?
or is the required setability 10% of the nominal value, like 47k +/-4.7k?
Depending on the requirement there may be an optimum solution with minimum contacts and maximum flexibility.
jd
Second try. I hope to use these gold-silver-copper hermetically sealed T0-5 relays for this project. I have used these same devices for more than 40 years. First at Ampex Research for servo control, and then many audio prototypes, over the years. They work, not perfectly, but compared to most solutions offered, pretty darn good.
Demian is a consultant for the same company as I am, and he knows what will go, and what will not. He and I have discussed this problem extensively with our employer already, so we are ahead of the general understanding of what is necessary.
Demian is a consultant for the same company as I am, and he knows what will go, and what will not. He and I have discussed this problem extensively with our employer already, so we are ahead of the general understanding of what is necessary.
Mercury relays are easy to get here. The company is called Meder Electronics.
Yes, contact resistance is quite high ( mercury is not a very good concuctor) but they can have very low thermo voltage. I heard many years ago from Kondo San of Audio Note Japan that he prefers a signal chain that uses only one kind of metal as conductor. In his case silver. You will not beleave it but he went so far as taking only silver from Italy that was held in stock for decades to "mature".
http://www.meder.com/fileadmin/meder/pdf/de/Produkte/Reedrelais/Relais_CRF_Serie_D.pdf
Yes, contact resistance is quite high ( mercury is not a very good concuctor) but they can have very low thermo voltage. I heard many years ago from Kondo San of Audio Note Japan that he prefers a signal chain that uses only one kind of metal as conductor. In his case silver. You will not beleave it but he went so far as taking only silver from Italy that was held in stock for decades to "mature".
http://www.meder.com/fileadmin/meder/pdf/de/Produkte/Reedrelais/Relais_CRF_Serie_D.pdf
John,
I am not trying to upset you, there are enough folks who seem to think that has value, I do not.
I would use 2 Pots with the lower range one on the N.O. contact as that side retains lower resistance than the N.C.
Yes, I would be picky about the control design because with a little work precise feedback is almost free. If you ground the wiper and use one side for the load, the other may be used for a position sensor! (Think 2 pots per for isolation!)
One of the projects on my bench is a speaker positioning system, almost the same as the RC volume control, just bigger motors! Last time I used a Bourns pot and the resolution and lifetime were excellent.
I am pretty sure you understand there are two ways to make a multiturn pot. Vishay does offer a bulk metal trimmer that is a screw driven wiper, as opposed to the classic Bourns' design which is a screw of resistance material with the wiper a follower.
But it is your design and as to the question of which pot to pick I have found that the shaft material has an effect on the Bourns' design types so these days I use the plastic shaft models!
Yes John,
You can use positive or negative feedback to change the apparent value of a loading resistor. Since you seem to be headed to using a 1K pot as the load. It can become a variable load from almost zero to 100K+ if you use 99% or more of positive feedback. Placing your gain range relays in that kind of system should also minimize their contribution.
Of course if you apply both positive and negative feedback to the pot the game becomes very interesting.
Simon, you are correct, however, I will leave it to you to do the project. Since my sources are all resistive and below 100 ohms, making a synthetic resistance (which I first did professionally, 40 years ago) is not necessary to actually improve the noise, but it might be a problem with audio quality.
John,
A good editor helps you present yourself better.
I think you may wish to expand upon the virtues of positive feedback. Although I first learned to use it from Popular Electronics in the 60's it is still rarely used. A discussion of how you can raise impedances and lower noise would be useful.
I have read both knots and in the first pile you mention B&K microphones. Today most audio measurements are done with a 1/4" microphone, but in the past 1" models were used due to s/n limits. Details of what happened should be interesting.
As to your motor driven load pot it seems you have all the design worked out.
I did a battery operated tube mic preamp a while back. Transformers were used on the input and output. A switchable adjustment was made for proper loading of the microphone. The gain was set low (35 db.) as the design intent was to allow for RFI free unclipped signal to go to an A/D for field recording. If I ever get around to building more I may steal the idea to make it a pot.
Oh by the way does anyone want to guess what the best selling most widely copied piece of sound equipment ever was?
- Status
- Not open for further replies.