[volt-nuts] [TekScopes] Re: OT: J511 current regulator diode (FET) temperature coefficient
David C. Partridge
david.partridge at perdrix.co.uk
Thu Jul 5 21:03:43 UTC 2012
I found your very helpful reply on the volt-nuts archive (reproduced here):
> I doubt the temperature coefficient of D717 matters in this circuit.
> Feedback is taken off of the output through TR705 or TR706 (x10 mode)
> back to the inverting input of IC705. The current source was used to
> prevent any significant shift in the operating point of TR703 for
> frequency compensation reasons because the switched cascode has
> voltage gain which is inside of the operation amplifier feedback loop.
> The drift in the cascode bias from D716 is irrelevant for the same
> Having voltage gain inside of the feedback does suggest another
> failure mode that could cause drift or noise in the output. Could
> IC705 or TR703 be oscillating? Maybe C714 dried out and has low
I have no idea why I never see some posts to that list!
C714 is suspect I agree, and I'll have to order in some 10uF tants. I did find that TR702 had a breakdown problem a while back and I replaced it temporarily with a 2N2907, pending the arrival of some hard to find BCY70 xistors - these are now here, but I've not gotten round to swapping that over.
I can definitely confirm that cooling or heating the CRD causes a large shift (millivolts) in the level at TP705. I've found no other component that causes that.
I need to check my other 7081 and a 7071 (identical) to see if they behave the same.
From: TekScopes at yahoogroups.com [mailto:TekScopes at yahoogroups.com] On Behalf Of David
Sent: 05 July 2012 19:06
To: TekScopes at yahoogroups.com
Subject: Re: [TekScopes] Re: OT: J511 current regulator diode (FET) temperature coefficient
Dave, did you see my reply over on volt-nuts at febo.com ?
It looks to me like D717 serves to bias the output of the cascode where its effect will be inside of the negative feedback loop. Its drift should be irrelevant if the circuit is working correctly.
I suspect they reason the designer did it that way is so that the operating current of the cascode would not change and because of the low compliance available. I expect the circuit is prone to oscillate because of the voltage gain inside of the feedback loop so the frequency and phase response had to be very carefully controlled.
I would have included a resistor between the emitters of TR702 and
TR703 to better control the gain although with the output shunted by at most 5KOhms, maybe that is not necessary.
On Thu, 5 Jul 2012 17:47:19 +0100, "David C. Partridge"
<david.partridge at perdrix.co.uk> wrote:
>It was the original part designed in, but I think the designer wasn't paying (enough) attention to the TC when he did that.
>This is affecting the output of the AC buffer circuit by over 1mV,
>which roughly means over 100uV error on AC readings on the 0.1V range
>as the meter warms up if you trim the offset when the meter is fully
>warm to be less than 100uV
>From: TekScopes at yahoogroups.com [mailto:TekScopes at yahoogroups.com] On
>Behalf Of Ed Breya
>Sent: 05 July 2012 16:34
>To: TekScopes at yahoogroups.com
>Subject: [TekScopes] Re: OT: J511 current regulator diode (FET)
>If that's the original part designed in, then it should be OK spec-wise, although you can always improve things.
>--- In TekScopes at yahoogroups.com, "David C. Partridge" <david.partridge at ...> wrote:
>> I'd already tested this in the dark too, as that problem was one of my first thoughts.
>> Looking at the datasheet for the J500 series diodes <http://www.datasheetcatalog.org/datasheet/vishay/70196.pdf>, I see that the J511 has a huge negative TC of 0.34% (34,000ppm) per degree C. Freeze spray on the diode or warmth from my finger confirms the problem - it is a horrid current source.
>> A Vishay CR470 would be a lot better at -1125ppm and even better would be a pair of CR240 current regulator diodes in parallel with a TC of zero according to the datasheet <http://www.datasheetcatalog.org/datasheet/vishay/70195.pdf>, but Vishay have discontinued those too.
>> Looks like I'll have to cobble up a bipolar current source as I think that will be a lot better.
>> David Partridge
>> -----Original Message-----
>> From: TekScopes at yahoogroups.com [mailto:TekScopes at yahoogroups.com] On
>> Behalf Of Ed Breya
>> Sent: 05 July 2012 03:30
>> To: TekScopes at yahoogroups.com
>> Subject: [TekScopes] Re: OT: J511 current regulator diode (FET)
>> temperature coefficient
>> Try the lid off-on experiments in the dark or with subdued light - you may have extra unwanted photoelectric bias currents from glass-packaged diodes in the circuit. You may be able to see that by shining a flashlight around at the circuitry. If you eliminate light effects, then you will be more sure of thermal effects.
Yahoo! Groups Links
<*> To visit your group on the web, go to:
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
(Yahoo! ID required)
<*> To change settings via email:
TekScopes-digest at yahoogroups.com
TekScopes-fullfeatured at yahoogroups.com
<*> To unsubscribe from this group, send an email to:
TekScopes-unsubscribe at yahoogroups.com
<*> Your use of Yahoo! Groups is subject to:
More information about the volt-nuts