Don't forget that fuse, HP changed the F2 fuse to 1/10 amp because
it was causing distortion for low measurements. These are listed in the
Manual Changes for MODEL 339A. Manual Part Number 00339-90001.
Also, HP discusses several changes with their photo-modules and some circuit changes with them. They upgraded them (some) and had a new part number.
I know in the long thread some of this was discussed not sure what
changes were recommended. Some of the photo-resistors in these
modules are the source of tracking errors.
From what I've researched on the photo-modules they get worse with
age.
Anyone source New, NOS, or better replacements? I would definitely be interested.
It's the SNR which needs to be improved. An input amplifier that delivers lower distortion at a higher level is what is needed. The noise floor is established in the early amplifier stages. The noise of amplifier stages down stream become less significant to the total noise. Most of the noise is just amplified from the input amplifier. As with any amplifier the noise is input referred. With an analyzer it's input amplifier referred.
The photocells are a big contributor to distortion. I tried some of the stock photocells but none of them worked. HP had a lot of parts made for them and I think the photocells are one of them. They are not used much in new design so what's out there is for repair use.
I haven't seen them used for anything other than daylight sensors in street and building lighting.
Just needed to prove to myself that the residual fund. wasnt in there as well. So, yes, its back to the noise issue. I have changed all opamps to a lower noise version - even the TL0xx types. I will tinker some more with the input amp.
[BTW - the THD+N of the osc as measured at the monitor output is <110dB.... just the analyzer wont go that low because of analyzer ?noise?] Which opamp did you use?
-RNM
Last edited:
LT1468. The voltage noise is only 2nV/rtHz better than what's in there. Current noise is of the essence here as the input Z is quite high, 100k ohms. With the attenuator before the input amp low Z source has little effect. With the Shibasoku there is little attenuation done before the input amp. Three ranges. That's all.
You might try OPA227 but the slew rate is low and I don't know if the required bandwidth will be realized. It has a current noise of 0.4nV/rtHz and voltage noise of 3nV/rtHz @ 1kHz. It is slow though.
The Silonex LDR's claim to have significantly less distortion. Figuring out the correct parts could be daunting without some guidelines on the existing parts. And Silonex was swallowed up by another company and the Audiohm parts seemed to fall of the supported product list. Advanced Photonix, Inc. | Optocouplers Silonex Optocoupler
I suppose one could removed one of the LDR, both are the same, and spec the thing.
I don't trust the data sheets on these things. I don't believe the manufacture can keep the tolerance tight on LDRs. The ones I tried were very close on the data sheet but neither would work in the 339A. The LED current to cell resistance has to be matched as well.
I don't trust the data sheets on these things. I don't believe the manufacture can keep the tolerance tight on LDRs. The ones I tried were very close on the data sheet but neither would work in the 339A. The LED current to cell resistance has to be matched as well.
David, from supplement, A2R22 change trimmer from 100 ohm to 200 ohm.
Add A2R43, resistor 90.9 ohm. For A2E1, which feeds the A2U10, pin 2 (-).
HP NOTE:
"Not all replacement photo-modules will work properly in instruments
which do not have t his modification."
On the following page there is a note regarding A3E1, "If it is necessary to change photo-module A3E1 be certain that A4R59 and A4R60 are the new values listed in this change."
A4R59 changes from 200 ohm t o 301 ohm. and A4R60
changes from 200 ohm to 100 ohm. The voltage level at the junction of R59
and R60 changes from +0.7 volt to +0.5 volts.
Here is the kicker for the A4R59, A4R60 change per HP,
"This change establishes a new reference from A4U6C to insure that the
'HI' frequency indicator is extinguished when the proper range is selected."
Does this imply if the indicator in on, when it shouldn't be, there are distortion
problems, tracking or other errors?
Also, somewhere in there are the changes to make your HF Filter response
better. Fix some problems with the Meter, etc. Some really good helpful
stuff and good explanations.
Okay time to go, baby girl is sick.
Last edited:
As she sleeps I found this:
Change the part number of A4U2 and A4U4 from 1820-0427 to 1826-0934.
A separate PN for Signetics part was established because PN 1820-0427
will no longer give the fundamental rejection required by the 339A.
Does this mean the original part has problems as it ages or that the
original supplier decides not to make the part any longer? Why not
say just replace the part?
Thanks Demian, funny thing all the "high-Performance" for audio LDR are all sold out. 🙁
How do we pronounce the name...Sy Lone EX? or See Lo NAY?
Change the part number of A4U2 and A4U4 from 1820-0427 to 1826-0934.
A separate PN for Signetics part was established because PN 1820-0427
will no longer give the fundamental rejection required by the 339A.
Does this mean the original part has problems as it ages or that the
original supplier decides not to make the part any longer? Why not
say just replace the part?
Thanks Demian, funny thing all the "high-Performance" for audio LDR are all sold out. 🙁
How do we pronounce the name...Sy Lone EX? or See Lo NAY?
Last edited:
Analyzer:
After checking the noise and the tuning on two different 339A's, I think it is the photo-modules which have changed more in one than the other; One I can tune null to show thd+n to -105dB (10dB better than spec) and the other with more extensive opamp changes, I cannot get below -95dB (spec). One 339 may have been ON for many more hours than the other.
I will be checking for NOS for that part. Maybe there is some compensation that can be done for the aging affects.... what ever that is.
THx-RNMarsh
Last edited:
I had the LT1468 at the input..... I found the LME49710 to be 2dB quieter and lower thd at the input location.... not much but I'll take it.
THx-RNMarsh
Last edited:
As I go back through my 339A I'll keep you advised about what I find.
I this case I found that I removed the A1C40 and A1C41 and replaced them
with High Quality BiPolar 100uf 50V caps.
I kept the original 47uf Tants. I dialed in the distortion using the HP339a's
own meter before and after the change. I lost 3dB from changing
back to the tants.
The reason for doing the changes is to undo all the mods I did to the HP339a
so that I can start from a clean slate. Sadly each time I undo a mod it takes
a big one. I've carefully undone the mods and with that latest undo
the lowest distortion I can do is -82 dB on the HP339a's own meter.
Richard, I'm afraid you are right, it is waterboarding.
I've now a selection of J110s, 2n4392s, and a couple of Vcr2n's
what is the best way to select the lowest noise JFET. There has
to be a simpler way than put it in the HP339a circuit and tweak it.
Even with installing single pin sockets for S, D, G, it is trick business
getting them to fit. With different packages turning around the plastic
case or installing with the metal case per HP.
Just looking for a good way to pre test and maybe pare down the field to
the top two or thee JFETS.
This is for the main Oscillator "variable resistor".
Cheers,
I this case I found that I removed the A1C40 and A1C41 and replaced them
with High Quality BiPolar 100uf 50V caps.
I kept the original 47uf Tants. I dialed in the distortion using the HP339a's
own meter before and after the change. I lost 3dB from changing
back to the tants.
The reason for doing the changes is to undo all the mods I did to the HP339a
so that I can start from a clean slate. Sadly each time I undo a mod it takes
a big one. I've carefully undone the mods and with that latest undo
the lowest distortion I can do is -82 dB on the HP339a's own meter.
Richard, I'm afraid you are right, it is waterboarding.
I've now a selection of J110s, 2n4392s, and a couple of Vcr2n's
what is the best way to select the lowest noise JFET. There has
to be a simpler way than put it in the HP339a circuit and tweak it.
Even with installing single pin sockets for S, D, G, it is trick business
getting them to fit. With different packages turning around the plastic
case or installing with the metal case per HP.
Just looking for a good way to pre test and maybe pare down the field to
the top two or thee JFETS.
This is for the main Oscillator "variable resistor".
Cheers,
The VCR2N is original selected part from NOS.... use it only. The bipolar cap changed for tant dont need to be replaced/removed for stock tant.
Its been so long ---- What IS and what WAS the problem you were having?
THx-RNMarsh
Its been so long ---- What IS and what WAS the problem you were having?
THx-RNMarsh
Last edited:
I continued to try and undo changes to the A4 board.
Until I had the opportunity to install the VCR2N, that
I received.
Demain had and open suggestion to use J110s and if I wanted
to see what difference they or any of them made to put single
pin sockets to see which JFET variable resistor on the A1 board.
So keeping along with that and your recent post to not change
caps to bipolar that have DC on them, I changed back the
two PS filter caps on the A1 board to the original Tant 47uf
caps A1C40 A1C41.
Trying to plot the my next course of actions.
If you have suggestions I would appreciate them.
Cheers,
Until I had the opportunity to install the VCR2N, that
I received.
Demain had and open suggestion to use J110s and if I wanted
to see what difference they or any of them made to put single
pin sockets to see which JFET variable resistor on the A1 board.
So keeping along with that and your recent post to not change
caps to bipolar that have DC on them, I changed back the
two PS filter caps on the A1 board to the original Tant 47uf
caps A1C40 A1C41.
Trying to plot the my next course of actions.
If you have suggestions I would appreciate them.
Cheers,
yes. ship it to me and I'll take a look at it.... dont want to see you being tortured any more nor drowning in this thing.
-RM
-RM
Sorry, Richard, I can't remember which part this is... more info? Curves for two different devices, with different resistances?
It clearly needs at least 300mVRMS to get to a linear control slope, but should remain below 1VRMS for best performance....
It clearly needs at least 300mVRMS to get to a linear control slope, but should remain below 1VRMS for best performance....
Assuming the flat lines at low signal levels are the result of instrumentation noise floors, are those curves telling me the ONLY nonlinearity in the LDR's is a third-order (cubic) factor?
Those plots look suspiciously like something generated from a simulation based on simplistic device models, at best . . . and perhaps simply a Matlab plot of what somebody thinks is SUPPOSED to be happening.
The plots in the atch App Note seem more credible to me.
Dale
p.s. - As I recall, the Silonex App Notes had a LOT of excellent engineering information, SPICE models, reference circuits, etc. They have disappeared from the Advanced Photonix web site. You can find some of them at http://www.cresttech.com.au/opto_index.html - grab copies while you still can.
Last edited:
Thanks Dale.
I looked or tried to look when Demian posted about some of the
photo-modules. Most of the links are dead.
I did find the one you posted and read it.
In the absence of good photo-modules what can we do instead?
Are there discreet circuits that can perform the same function
with little hassle?
Can something be made on a small board
that might work better with less distortion?
Surely there are plenty of big brains around here
that could do it...I'd take a stab at it but I seem to
lack the knowledge, skills, training, and equipment
to even revive my HP339a.
Richard was kind enough to help me out and end my suffering
and hopefully will find my analyzer at his doorstep any day now.
Demian spent some time personally to help me sort out some
details and identified some problem areas that I couldn't identify
what the underlying problem was...neither could he.
David and RichEEM have also spend countless hours providing
their thoughts and guidance too.
I have learned what not to do and it opened my eyes and mind.
I can also get sidetracked which presents it own set of challenges.
If anyone has a spare HP339a or or know where I can find the following
knobs:
1. Distortion Input Range knob and skirt.
2. Fine tune knob (Oscillator Level).
Much appreciated.
I looked or tried to look when Demian posted about some of the
photo-modules. Most of the links are dead.
I did find the one you posted and read it.
In the absence of good photo-modules what can we do instead?
Are there discreet circuits that can perform the same function
with little hassle?
Can something be made on a small board
that might work better with less distortion?
Surely there are plenty of big brains around here
that could do it...I'd take a stab at it but I seem to
lack the knowledge, skills, training, and equipment
to even revive my HP339a.
Richard was kind enough to help me out and end my suffering
and hopefully will find my analyzer at his doorstep any day now.
Demian spent some time personally to help me sort out some
details and identified some problem areas that I couldn't identify
what the underlying problem was...neither could he.
David and RichEEM have also spend countless hours providing
their thoughts and guidance too.
I have learned what not to do and it opened my eyes and mind.
I can also get sidetracked which presents it own set of challenges.
If anyone has a spare HP339a or or know where I can find the following
knobs:
1. Distortion Input Range knob and skirt.
2. Fine tune knob (Oscillator Level).
Much appreciated.