I have an old Heathkit AA-1 where some of the resistors apparently corroded from moisture from the looks of them(swollen ends and greyish film on sleeve) and are open. I am considering how to fabricate new ones and was wondering if anyone has done this in the past. I believe the circuit is low current.

I've attached the schematic below and the divider in question is the lower left switch. Note the values are all similar except the decade.

 

The E96 series of values (1% tolerance) gets pretty close, e.g. 216, 316, 681.

 

Why would you want to fabricate a resistor instead of just purchasing commercially manufactured ones?

As for hitting the values in the schematic, I doubt you can. For instance, a value of 68.38Ω implies a tolerance on the order of 0.01%. Even if you trim a resistor to that value, it is unlikely that it will stay that value for long as the temperature changes and as it ages. Notice that all the resistors are in decade sets, so I'd recommend getting 0.1% resistors that are the closest to these unless you can determine what the governing relationship is to get a set of nominal values based on obtainable resistors.

 

How to fabricate new ones? A four digit label indicates 0.1% resistors. Look to buy 0.1% resistors or get a really good ohm meter and play series/parallel with some 1% resistors.

 

The E96 series of values (1% tolerance) gets pretty close, e.g. 216, 316, 681.

I'm showing 215 instead of 216.

 

Hello,

You can calculate odd resistor values with SgtWookies calculator shown in this post:
http://forum.allaboutcircuits.com/t...ps-and-tricks-thread.12878/page-3#post-387305

Bertus

 

Hello,

You can calculate odd resistor values with SgtWookies calculator shown in this post:
http://forum.allaboutcircuits.com/t...ps-and-tricks-thread.12878/page-3#post-387305

Bertus

I was looking at a parallel calculator I found thinking that some SMD resistors on a small piece of copper clad with a gap would be the way to go. I found I could get pretty much spot on with 3 resistances. The SMD items were rated 1% 1/4w at 25ppm/C. Just wondering if there was something already worked out. Thanks for the input.

 

There is also a utility on the TI site. Unfortunately, it does not appear to be working right now: http://www.ti.com/tool/parallelrescalc

John

 

How close are you going to try to hit those nominal values? Remember that with 1% resistors, you are already working with a much wider window than the nominal values imply. For instance, let's say that you are focusing on the 6838 Ω resistor. If you start off with a 6810 Ω resistor then you might think that you only need to add another 28 Ω resistor in series. But remember that the 6810 Ω resistor has a tolerance of ±68 Ω, well over twice the nominal value of the resistor that you would think you need to add such that there is a fair chance that the actual resistance is already too large and that adding the 28 Ω resistor is only making it worse. Also, at 25ppm/°C, it only takes a 6°C temperature change to result in a 1 Ω change in the resistance, so trying to hit 6838 Ω exactly just doesn't make a lot of sense.

 

The values are avaialble on Digikey - just not in stock and a 5000 unit order to get them stocked.

Maybe they can be found elsewhere - Heathkit repair/rebuild kits on ebay anywhere?

Seems to be a Vishay - 0.1% tolerance series
CMF 55 (1/8 watt)
CMF60 (1/4 watt)

 

I was looking at a parallel calculator I found thinking that some SMD resistors on a small piece of copper clad with a gap would be the way to go. I found I could get pretty much spot on with 3 resistances. The SMD items were rated 1% 1/4w at 25ppm/C. Just wondering if there was something already worked out. Thanks for the input.

When I want parallel, I just stack the smd devices (e.g., capacitors and resistors) and solder up the edge. Doesn't take any extra board space that way.

John

 

I'm showing 215 instead of 216.

Just shows you can't rely on the Internet for accuracy/consistency .

 

Just shows you can't rely on the Internet for accuracy/consistency .

You are so right about that. So I'm curious, where did you find the reference that had 216? I checked three and they all had 215, but that sure doesn't mean that even those three sites actually agreed on all the values.

EDIT: I had inadvertently swapped the 215 and 216 above. All of the references I found give 215 as members of the E48, E96, and E192 sequences.

 

You checked three references for a typo?
That explains a lot.

 

I seem to remember something way back in my memory bank about taking a carbon resistor and grinding the carbon down to get an odd value, cutting a V in the center. If that actually works a modern moto tool would make the process fairly painless. Not sure what it would do to the wattage rating and maybe my memory is all wet.

 

A lot of carbon film and metal film resistors are already spiral cut for calibration. Very easy to ruin them. Don't do that with the kind of resistors that are made for 1% ratings.

 

You checked three references for a typo?
That explains a lot.

I wanted to have some confidence that the value I had found in the one reference I used was correct compared to the other one since I don't have the E48 and higher tables memorized. So I checked another reference and it agreed with the first. Then, since I already had the Google results there, it was a simple matter to check a third.

And, yes, I tend to at least try to chase down inconsistencies in anything I do or am involved with.

 

I seem to remember something way back in my memory bank about taking a carbon resistor and grinding the carbon down to get an odd value, cutting a V in the center. If that actually works a modern moto tool would make the process fairly painless. Not sure what it would do to the wattage rating and maybe my memory is all wet.

I won't recommend that with modern resistors, even bulk carbon composition (which are getting harder to come by anyway). Aside from the fact that many, particularly carbon film, resistors are laser trimmed and it would be easy to ruin the resistor that way, you are also breeching the protective coating and asking for problems down the road.

 

@ WBahn

So I'm curious, where did you find the reference that had 215?

I didn't; you did (post #5)

 

You can also use the formula

R = 10 ^(i/E)
where E is the number of steps in the series and belongs to {12, 24, 48, 96,...}
and i belongs to {0,1,...,E}

Depending on how you view the roundoff process, either 215 or 216 is a reasonable choice.
You get 215 if you truncate, or round down, or round to the closest integer.
You get 216 if you round up.