Amber 3501A Audio Analyzer Needs Refurb and Cal any info would be helpful.
Hello
Does anybody know of somebody who could Refurb and cal. a Amber 3501A Audio Analyzer. Also if somebody has refurb there own any insights would be appreciated. If I can't find a tech I will do it myself.
Please PM me with any info or post in the thread.
Thanks in Advance.
bOTL
Hello
Does anybody know of somebody who could Refurb and cal. a Amber 3501A Audio Analyzer. Also if somebody has refurb there own any insights would be appreciated. If I can't find a tech I will do it myself.
Please PM me with any info or post in the thread.
Thanks in Advance.
bOTL
Last edited:
Please don't bump more than once a week per forum policy.
That said I have a 3501A that I refurbished about 8 years ago.
No time right this moment, but tell me a little more about the condition of the unit. With care if you can solder and do some simple diagnostic work you can do the work yourself.
That said I have a 3501A that I refurbished about 8 years ago.
No time right this moment, but tell me a little more about the condition of the unit. With care if you can solder and do some simple diagnostic work you can do the work yourself.
Hi bOTL,
I picked up an Amber 3501 about 4 - 5 months back. I do not know of anyone to do the work your wanting however you could do it yourself. I started doing a bit of refurbishing work on mine over the past few months, so far I've replaced the tantalum caps and a number of the multitude of 10/40 e-caps. I also removed the freq potentiometer stack from the board and flushed/cleaned the 4 pots.
I experimented and installed some LME 49710's in place of some of the 5534's in the oscillator section and I was able to tune the 2nd harmonic down to about 10 - 12dB lower level than with the 5534's. In fact the 2nd harmonic dropped into the noise floor of my QA400. However it seems that the tune slowly drifts on me, if I go back after a couple of days the second harmonic has crept back up. The single turn adjustment pots are ridiculously touchy and I think that they may be drifting, I'll probably go back in and put in some multi-turn pots in their places. If you decide to tackle doing this work yourself I'd recommend replacing these single turn adjustment pots.
The manual that is available online has a good amount of information in it about setting the unit up and calibrating it.
I picked up an Amber 3501 about 4 - 5 months back. I do not know of anyone to do the work your wanting however you could do it yourself. I started doing a bit of refurbishing work on mine over the past few months, so far I've replaced the tantalum caps and a number of the multitude of 10/40 e-caps. I also removed the freq potentiometer stack from the board and flushed/cleaned the 4 pots.
I experimented and installed some LME 49710's in place of some of the 5534's in the oscillator section and I was able to tune the 2nd harmonic down to about 10 - 12dB lower level than with the 5534's. In fact the 2nd harmonic dropped into the noise floor of my QA400. However it seems that the tune slowly drifts on me, if I go back after a couple of days the second harmonic has crept back up. The single turn adjustment pots are ridiculously touchy and I think that they may be drifting, I'll probably go back in and put in some multi-turn pots in their places. If you decide to tackle doing this work yourself I'd recommend replacing these single turn adjustment pots.
The manual that is available online has a good amount of information in it about setting the unit up and calibrating it.
Assuming you don't already have a manual, here is a link to the 3501 service manual. There are a few things that differ; the 3501A does not support battery operation or the IMD option, most of the rest of the circuitry outside of the PSU are identical.
http://www.richardhess.com/manuals/Amber3501.pdf
http://www.richardhess.com/manuals/Amber3501.pdf
I have a 3501 with the battery option and some other options. It seems really nice but not the easiest to work on. Its mostly working but was a distraction so it's just sitting. (I have quite a collection of distortion analyzers.) I just bought a 5500 which may fit my needs better. Where can I get a manual for it? I have tried the easy options.
PM sent.
Glad to see the forum back up. I had to deal with a similar dbf corruption about a month back.
I'm starting a separate thread on the Amber 5500. My 3501 is working but needs a tuneup to get to specified performance. if someone does go through one a tutorial or sequence of test/troubleshooting would be helpful. They are remarkable little boxes. I suspect a sequence of cleaning contacts and calibration adjustments would be all that is necessary. Worst case may be replacing the optocoupler. The Luna optocouplers may be a significant improvement https://www.digikey.com/products/en?mpart=NSL-32SR3&v=209 The original part is a Clarex CML8500. On resistance is spec'ed at 5K Ohms max. So almost any current variation would work.
Demian, I know that you're fairly thorough on the refurb and tweaking of your test gear so I'm very interested in how the 5500 works out for you. Based on the low readings I'm able to get from my 3501's analyzer (.00045%) I suspect the last gen 5500 will be an even better piece.
There are a few quirks with the 3501. I have one of the last versions (1988 mfg date) of this analyzer with the internal switching PS and it runs hot. After an hour or so the case of the unit is quite warm. I noticed that it is quietest and measures best when its first turned on and cold. If I let it sit for a while it gets a slight bit noisy on the measurements and then settles down again a bit later, like after 45min, but the measured readings are ever so slightly worse than when the unit was cold. I have a wall mounted fan so I now just direct the fan towards my stack of gear and this seems to take care of the heat instability issue. I noticed that the Amber 5500 has a rear cooling fan, I guess Amber decided the heat was a big issue and needed to be corrected.
I've also noticed that the 3501 is mildly unstable when the freq tuning knob is set in the middle of the range. If you get to either end of the range, the bottom or the top, then the distortion measurements improve drastically. As an example if I set the knob to say 9.8 X 100 the best distortion I can measure from my modded 339A oscillator is about .0008% and the needle wavers about .0001% or more. If I rotate the knob up to just past 10.0 X 100 where all of the potentiometers have reached full travel the distortion drops dramatically to .00045% and stability is also much improved. So potentiometer alignment is very important. There's four pots in the 3501 tuning stack and I pulled them and cleaned them when I was recapping the board. I measured the resistance of the pots and they were originally selected as somewhat closely matched pairs based on my measurements. I made sure to reinstall them into their same locations as well. My thought was that the noise I was seeing was due to the pots needing to be cleaned, however that turned out to not be the case as after reinstallation they are as bad if not more so than before cleaning. Its got me wondering how the unit would respond to a fixed set of 1% resistors set up in 4 steps of 1, 2, 5 & 10 to get a reasonable freq spread and using small relays to switch them in/out.
BTW I picked up a spare Clairex CLM8500 off of an ebay seller that is selling originals in the OEM blister pack.
There are a few quirks with the 3501. I have one of the last versions (1988 mfg date) of this analyzer with the internal switching PS and it runs hot. After an hour or so the case of the unit is quite warm. I noticed that it is quietest and measures best when its first turned on and cold. If I let it sit for a while it gets a slight bit noisy on the measurements and then settles down again a bit later, like after 45min, but the measured readings are ever so slightly worse than when the unit was cold. I have a wall mounted fan so I now just direct the fan towards my stack of gear and this seems to take care of the heat instability issue. I noticed that the Amber 5500 has a rear cooling fan, I guess Amber decided the heat was a big issue and needed to be corrected.
I've also noticed that the 3501 is mildly unstable when the freq tuning knob is set in the middle of the range. If you get to either end of the range, the bottom or the top, then the distortion measurements improve drastically. As an example if I set the knob to say 9.8 X 100 the best distortion I can measure from my modded 339A oscillator is about .0008% and the needle wavers about .0001% or more. If I rotate the knob up to just past 10.0 X 100 where all of the potentiometers have reached full travel the distortion drops dramatically to .00045% and stability is also much improved. So potentiometer alignment is very important. There's four pots in the 3501 tuning stack and I pulled them and cleaned them when I was recapping the board. I measured the resistance of the pots and they were originally selected as somewhat closely matched pairs based on my measurements. I made sure to reinstall them into their same locations as well. My thought was that the noise I was seeing was due to the pots needing to be cleaned, however that turned out to not be the case as after reinstallation they are as bad if not more so than before cleaning. Its got me wondering how the unit would respond to a fixed set of 1% resistors set up in 4 steps of 1, 2, 5 & 10 to get a reasonable freq spread and using small relays to switch them in/out.
BTW I picked up a spare Clairex CLM8500 off of an ebay seller that is selling originals in the OEM blister pack.
I saw those Clarex parts as well. The Luna started with a different owner some years ago and were supposed to be significantly lower distortion at a given setting/level. That would make them worth exploring.
I think there is something less than right on the input circuit of the 5500. I noticed that the generator distortion is really low until its connected to the input of the analyzer and then it jumps up. My guess is some part in the input protection is not working right. This is one of my standard tests and all too often input nonlinearity shows up, especially on higher impedance sources. The Shibasoku is good for this test. The input circuit of the Amber should also be good but maybe something is not obvious.
I think there is something less than right on the input circuit of the 5500. I noticed that the generator distortion is really low until its connected to the input of the analyzer and then it jumps up. My guess is some part in the input protection is not working right. This is one of my standard tests and all too often input nonlinearity shows up, especially on higher impedance sources. The Shibasoku is good for this test. The input circuit of the Amber should also be good but maybe something is not obvious.
I installed a few LME49710's, that's all I had on hand, into the oscillator section in place of a few of the 5534's. I wasn't able to make a huge dent in the loopback performance. I was able to tweak the 2H down to a lower level than with the 5534's while watching the output on my QA400 but from what I remember I was still dealing with a large 3H.
I'd love to get the 3501 oscil to match the performance of its analyzer section. The best I've been able to get the 3501 oscillator is about .00075% at 1kHz. My modded 339A oscillator is now measuring .00037% when fed into the 3501 analyzer. This reading is only possible with the potentiometers at full travel 10 X 100 as I stated in post #14. I'd probably sell the 339A if I could get the oscillator section performing a bit better on the 3501. I need to go back and dabble with it a bit more. I'm just waiting on some cooler weather as its miserably hot in my shop in the evening when I get home.
I bought a couple of the Luna optocouplers to try out, per your previous advice. They were cheap enough and I had to order something from Digikey so I thought what the hell lets give this a try...
I'd love to get the 3501 oscil to match the performance of its analyzer section. The best I've been able to get the 3501 oscillator is about .00075% at 1kHz. My modded 339A oscillator is now measuring .00037% when fed into the 3501 analyzer. This reading is only possible with the potentiometers at full travel 10 X 100 as I stated in post #14. I'd probably sell the 339A if I could get the oscillator section performing a bit better on the 3501. I need to go back and dabble with it a bit more. I'm just waiting on some cooler weather as its miserably hot in my shop in the evening when I get home.
I bought a couple of the Luna optocouplers to try out, per your previous advice. They were cheap enough and I had to order something from Digikey so I thought what the hell lets give this a try...
Both the oscillator and analyzer are less that wonderful on this one. It doesn't match the manual on the web in some ways. I figured out the controls for each but they both just don't quite get there. There are so many opamps one could go nuts replacing all of them however that can be counterproductive. With the options on this one getting to the I/O board while its operating seems unlikely since its under a layer of sheet metal. I'll order 20 LME49710's and start shotgunning. It will be interesting to see what I find.
Following, I overhauled my ex Canadian Navy 3501A about 8 yrs ago after a relay failed and one of those tantalum drop capacitors caught on fire.
I replaced them with aluminum electrolytics (low impedance type) and noted a very slight increase in measured thd in loop back testing at 1kHz.
Tonight in loop back at 1Vrms out I was able to get no better than 0.0013% at 1kHz with 30kHz LPF.
I've owned this unit since 1999 (spring/summer?). I guess maybe I need to look at some improvements when I have time. Perhaps start by replacing those electrolytics I mentioned with Oscons or similar..
I replaced them with aluminum electrolytics (low impedance type) and noted a very slight increase in measured thd in loop back testing at 1kHz.
Tonight in loop back at 1Vrms out I was able to get no better than 0.0013% at 1kHz with 30kHz LPF.
I've owned this unit since 1999 (spring/summer?). I guess maybe I need to look at some improvements when I have time. Perhaps start by replacing those electrolytics I mentioned with Oscons or similar..
Demian,
At least the IC's are all socketed so its easy to make a change.
My 3501 had a few differences between the web manual and what I saw as well.
I went back and looked at mine. On the analyzer side I have 49710's at U2008, 2011 and 2017 in the notch filter. On the input attenuator pcb I have a 49710 at U1030. It seems to me that U2008 in the notch summing amp made the biggest change however I was following the shotgun method of scientific experiment so I do not have any specific results for each socket. I picked up another handful of 49710's with my last parts order so I could do some more experimenting.
At least the IC's are all socketed so its easy to make a change.
My 3501 had a few differences between the web manual and what I saw as well.
I went back and looked at mine. On the analyzer side I have 49710's at U2008, 2011 and 2017 in the notch filter. On the input attenuator pcb I have a 49710 at U1030. It seems to me that U2008 in the notch summing amp made the biggest change however I was following the shotgun method of scientific experiment so I do not have any specific results for each socket. I picked up another handful of 49710's with my last parts order so I could do some more experimenting.