John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
In order to keep things in order. I STILL don't have any effective suggestions as to ADDING to the VARIABLE pot to extend its range, either up for less loading, or INCREASING its resolution. We do have, on the way, a P3 10 turn 10K hybrid multistage pot that should arrive within the week. I suspect that its resolution will be too tight for and easy 130 -138 ohms adjustment, for example. I know that people have worked on this for years, but I never bothered, because I have almost always used a variable pot,exclusively. Any suggestions out there?
 
AX tech editor
Joined 2002
Paid Member
In order to keep things in order. I STILL don't have any effective suggestions as to ADDING to the VARIABLE pot to extend its range, either up for less loading, or INCREASING its resolution. We do have, on the way, a P3 10 turn 10K hybrid multistage pot that should arrive within the week. I suspect that its resolution will be too tight for and easy 130 -138 ohms adjustment, for example. I know that people have worked on this for years, but I never bothered, because I have almost always used a variable pot,exclusively. Any suggestions out there?

Well, some of us tried, some tried to get an idea of what range you needed, but nothing was forthcoming. Then you said that the TO5 G-o-G relay was non-negotiable. Next you say you want to explore alternative switching elements. Then Simon suggested a switching approach and you told him, go do your own project. First you tell us the 1k pot is not negotiable. Now all of a sudden it's a 10k pot. You have an interesting engineering approach, I grant you that ;)

jd
 
Ed Simon, would you please answer your own question?

John,

You guess was pretty good, but there are lots of others who might want to try.

As to a 10K sample pot, about all you can do is parallel a 1K resistor, Then switch in or out a 100 ohm unit. A 1K trim pot may have about 1 ohm of termination resistance so you might want to put a few ohms in that leg. It should help low value tracking.
 
It is just a test pot, we probably will actually purchase a 1K pot. I regret if I have made the problem that I was hoping to address a little scattered. Here it is again: At this time (subject to change by my employer as it has already changed over the weeks) it is now 10 ohms-1K variable, then switched up to 100K. I am still looking for optimum switching from 1K to 100K. Any suggestions?
 
AX tech editor
Joined 2002
Paid Member
It is just a test pot, we probably will actually purchase a 1K pot. I regret if I have made the problem that I was hoping to address a little scattered. Here it is again: At this time (subject to change by my employer as it has already changed over the weeks) it is now 10 ohms-1K variable, then switched up to 100K. I am still looking for optimum switching from 1K to 100K. Any suggestions?

I don't know why your employer takes design decisions - isn't that why he hires people like you?
Anyway, from an engineering standpoint a 10t 10k appears a better choice. At 1k per turn there's enough resolution in the lower 100's area, and you could still use an (anti) log fake resistor to improve resolution even more. 10k should allow you to get away with less switching and less relays or whatever you want to use for switching. You could have one range 10k-10 ohms (with fake antilog or log law) and then 100k-10k. Are you foreseeing to divide that 100k-10k in pieces with switches or somehow with the same pot?

jd
 
Last edited:
How about something like this? Different combinations of relay switching allow smooth adjustment from 0 to 80k.

Perhaps a bit of overlap would be good e.g. 4 relays with fixed resistors of 10k, 18k, 33k and 56k.
 

Attachments

  • relays.GIF
    relays.GIF
    1.6 KB · Views: 257
Jan, I have previously stated that the 'sweet spot' has to be in the 130 ohm area. 10K is fairly far away, that is why I am NOW considering 1K or 2K as a FINAL adjustable pot.
I would like to point out that I have used just about every pot that is normally available, over the years, and I have samples in my lab. Including P&G rotary, linear, and military, Alps, both Black and Blue, TKD, Bourns, AB, etc.
I even have a fair number of wirewound 10K pots. I have tested them already. My best hope is hybrid and probably a lower value than 10K.
Spot on Godfrey! That is just the sort of input I was hoping for.
 
I don't know why your employer takes design decisions - isn't that why he hires people like you?
Anyway, from an engineering standpoint a 10t 10k appears a better choice. At 1k per turn there's enough resolution in the lower 100's area, and you could still use an (anti) log fake resistor to improve resolution even more. 10k should allow you to get away with less switching and less relays or whatever you want to use for switching. You could have one range 10k-10 ohms (with fake antilog or log law) and then 100k-10k. Are you foreseeing to divide that 100k-10k in pieces with switches or somehow with the same pot?

jd

Jan,

You forgot rule #2. "When the boss is wrong refer to rule #1!"

Godfrey,

Yes that is one of the two basic circuits, and you got right that the resistors need to be a bit low to allow for tolerances. In the case of a 1K pot with an end resistor the resistors should be 988 ohms, 1.91k, 3.75K, 7.39K, 14.3k, 28K and 54.9K. The other method is an R/2R ladder see

Resistor ladder - Wikipedia, the free encyclopedia

The actual issue is how to minimize the use of relays. Ideally only one relay should be closed. So the serial method with more relays may be a good choice, except that relays are expensive.
 
Cost is not so important, minimum number of series relays is important.
Hi John

It is a pity that it is impractical to use two separate pots (for high range and low range). That would allow only one relay contact to be in circuit at a given time. It would also be easier to use with only 2 switch positions.

I understand the difficulty of having special pot/motor combinations custom-built, but would it not be possible to use two completely separate motorized pots?

Regards - Godfrey
 
In the case of a 1K pot with an end resistor the resistors should be 988 ohms, 1.91k, 3.75K, 7.39K, 14.3k, 28K and 54.9K.
Hi Simon

The problem with using this kind of approach with a 1k pot is that you end up with about 100 different switch settings.

Just sweeping e.g. from 50k to 60k will involve going through 10 combinations of series resistor. There will be much clattering of relays - not ideal while one is trying to tune by ear!

Cheers - Godfrey
 
Hi Simon

The problem with using this kind of approach with a 1k pot is that you end up with about 100 different switch settings.

Just sweeping e.g. from 50k to 60k will involve going through 10 combinations of series resistor. There will be much clattering of relays - not ideal while one is trying to tune by ear!

Cheers - Godfrey

Depending on the complexity of the controller more relays can be used to minimize the number of switches being used as shunts. I don't think the miniature relays being used are as noisy as you expect. Inside a solid case or two they should be well below the problem level.

Somewhere around 15 relays will limit you to mostly 2 switches sometime 3.
 
Last edited:
Status
Not open for further replies.