<VV> Thermister Swap?

Mon Sep 8 00:03:31 EDT 2014

A couple of things struck me, reading the last few posts on the thermistor. 
 I have a copy of Herb's characteristic curve for the sender. Herb isn't 
around  to ask his sampling methods, but it would be interesting to know the  
original design curve for this unit, not just a single sample. At least a  
few more units through test would provide more confidence that this 
represents  the designed goal. The same is really true in the instrument area. In 
addition  to the input voltage - and it would be nice to see what the real 
world provides  at a Corsa dash plug under running conditions - the level of 
grounding  of the instrument panel assembly (a Corvair 'issue" over the years), 
would be  nice to instrument. This is all good info. - Seth  

In a message dated 9/7/2014 6:03:10 P.M. Pacific Daylight Time,  
virtualvairs at corvair.org writes:

I've  done some more research on the Corsa thermistor, stock gauge and  the
Rochester thermistor and I may have been WRONG in my earlier posting  saying
that the Rochester thermistor with a fixed resistor wouldn't  work.

The major problem I have is that there's little data on the stock  GM/AC
Delco thermistor.  There's one plot made by Herb Berkman back in  1974 of
resistance versus temperature.  There's no data I can find  regarding the
Corsa dash gauge; I think we've all assumed that GM designed  the gauge to
match the thermistor characteristic curve.

Since there  was little information on the thermistor, and I don't have a
handful to  make new measurements from, I decided to back into the problem
and see what  the Corsa gauge "expected" to see at various temperature
readings.  At  least I had three gauges to check and see if they were
consistent.   Basically I hooked up a Corsa head temp gauge and substituted
a variable  resistor and measured the the resistance needed to make the
gauge indicate  various temperatures.

I was shocked by my results!  If Mr.  Berkman's data are correct, the stock
Corsa setup is about correct at only  200 degs F cylinder head temp.  After
that, it reads progressively  LOW.  Based upon my measurements of the gauge
response and Mr.  Berkman's data, I'd estimate that at a real head temp of
500 F, the Corsa  gauge is indicating about 350 F.  That's a scary error!
I'd rather it  read too high than too low!  Basically the stock Corsa gauge
and the  AC/Delco thermistor don't match each.

Regarding the Rochester  thermistor, the only specs I can find were
published in a Robinson  Helicopter instruments maintenance manual.  (I
can't find any specs on  the Rochester web page; perhaps I just don't know
where to look...)   If the Robinson data is correct, then adding a 60 ohm
resistor in series  with the Rochester thermistor would make the stock Corsa
gauge work pretty  well.  It would be just about right on at 500 degrees and
track quite  well at lower temperatures -- probably well within your ability
to read it  accurately.  (Above 500 degrees, the Corsa gauge with  Rochester
thermistor + 60 ohm resistor would again read lower than the  actual temp.
But I would be backing off by the time my car's heads hit 500  F...)

So I now need to get the real specs on the Rochester thermistor  or actually
buy one and measure it's resistance vs temperature  curve.

If anyone want's to see my data and analysis, drop me a note and  I'll
forward a copy of the spreadsheet with graphs.

Jim  Simpson
Group Corvair

On Fri, Sep 5, 2014 at 4:33 PM, MarK Durham  <62vair at gmail.com> wrote:

> Jim is exactly right. It would be  accurate at one point, and less 
accurate
> progressively along the scale  going up or down from that point. That is 
why
> I said on the EFI, it  took an approximate reading to fool the computer in
> switching to  normal temp mode form cold start mode and a resistor did 
that.
> The  computer only needed to see an approximate value within a  range.
>
> A guage needs to be more accurate than  that.
>
> Guys, its this type of thing that is happening to our  aging cars that
> prompted me to suggest installing a new movement in  the dash to replace 
the
> old one, (and you keep the original look) and  to install a matching new
> thermistor in the engine, so you have  correct reading gauges. If its CHT,
> VDO makes a 90 degree sweep guage  that could be a direct replacement 
with a
> VDO probe. Then your CHT  would be good for longer than most of us would
> have the cars! VDO also  does transmission temp or oil temp gauges.
>  Regards
>
>
>
> Mark Durham
> Hauser,  Idaho
> 62 Monza coupe Red/Red 4 speed
>
>
> On Fri,  Sep 5, 2014 at 12:32 PM, Jim Simpson via VirtualVairs <
>  virtualvairs at corvair.org> wrote:
>
>> The short answer is  no, it won't work.  The longer answer is that yes, 
you
>> can  put a suitable resistor in series with the Rochester thermistor  and
>> get
>> a correct reading at one (1)  temperature.  Pick the temperature you want
>> to
>> be  "accurate" and you could find a resistor that would give you that  
one
>> reading.
>>
>> The problem is that the  Rochester and GM thermistors have different
>> temperature vs  resistance curves.  Even with the addition of a resistor,
>> the  curves will not overlay.   All you can do is make them cross  each
>> other
>> at some point.
>>
>> If  you want to use the Rochester thermistor, you will have to make  (or
>> have
>> made) a black box to translate from it's  response curve to the gauge
>> response curve.  In principle,  not a hard thing to do.  But it would 
take
>> a
>>  little time and tinkering.
>>
>> I've been thinking about it  a little and may give it a try in my spare
>> time
>> in  order to refresh my electronics skills.  But don't hold your  breath
>> waiting for me to get it done...
>>
>>  Jim Simpson
>> Group Corvair
>>   _______________________________________________
>> This message was  sent by the VirtualVairs mailing list, all copyrights
>> are the  property
>> of the writer, please attribute properly. For help,  mailto:
>> vv-help at corvair.org
>> This list sponsored by the  Corvair Society of America,
>> http://www.corvair.org/
>>  Post messages to: VirtualVairs at corvair.org
>> Change your  options:
>>  http://www.vv.corvair.org/mailman/options/virtualvairs
>>   _______________________________________________
>>
>
>
_______________________________________________
This  message was sent by the VirtualVairs mailing list, all copyrights are 
the  property
of the writer, please attribute properly. For help,  
mailto:vv-help at corvair.org
This list sponsored by the Corvair Society of  America, 
http://www.corvair.org/
Post messages to:  VirtualVairs at corvair.org
Change your options:  http://www.vv.corvair.org/mailman/options/virtualvairs
_______________________________________________