-
Was hoping someone smarter than me could help out with this. Thanks in advanced!
So I pickup up a "Slightly" used Valhalla 2701C marked for "parts" at my local surplus store. As the title says it kinda works. I've read a few other posts on this model but I think I may need more help.
First off the unit does power up and I get voltage output! The output seems to correspond with the voltage displayed but its unstable. Description below. My diagnosing skills may not be enough to fix this thing, but I do like a challenge.
Thanks to user lowimpedance for the schematic on this unit!
When the unit powers up it runs through a short welcome message. With power on relay RLJ clicks constantly about once per second. The front range selection switch seems to work and I do get a corresponding voltage. But due to the relay clicking there is a spike in the output. I believe the spike is causing an instability in the control hardware also. The unit "locks up" randomly and the front controls stop working. This happens randomly sometimes after a few minutes powered on.
https://drive.google.com/open?id=1Fy1EsrgRXZ0IwHGnnXPGn2uLaU0tUNr- 2V selected
https://drive.google.com/open?id=1oAW_TKlFwnM58WB2yDHnRtEvvg0mw90v 120v selected
Here's a short video of the behavior.
The switch was broken but I fixed it later. You can see the unit "lock up" I think due to the spikes. After a power cycle it kinda works.
https://drive.google.com/open?id=1uMT1Kxave_T29GaW5Qx280OlXdqrFF48
Showing the spike on the oscilloscope. Notice the unit is in standby.
https://drive.google.com/open?id=1TBeBGnTKty16fTEja68fZVGEd4bIZGn4
This is a relay transistor TR17 (2N4402 PNP). It looks like it may have been removed/replaced. I removed and tested it. All seems normal so I reinstalled the part mainly because I don't have any suitable replacements on hand. TR17 is connected to IC5 which is a (icl7650scpd 2MHz, Super Chopper-Stabilized Operational Amplifier). The voltage regulators seem good and the IC is powered. I followed the output to D17 and C17 to the base of TR17. C17 has about 1.2v across it while the unit is powered. I did replace with a newer 35v 1000uf cap but didn't change the behavior. Here's a picture before I removed the transistor.
https://drive.google.com/open?id=14LRWMnHSjZCIqdzDUmO7mqLcTVcqJ3ug
A few of the HV transistors look like they have been replaced. I've probed around and there is defiantly HV there. Around 400v at each of the tabs of the HV transistors to ground.
https://drive.google.com/open?id=19DNu80zaExc27fCGdAFBw6YpdrdinMsh
Close up of the board
https://drive.google.com/open?id=10I4o940mJqTTUKifjPa7m2jNztfxcv84
There were no obviously burned or damaged parts. I've probed all the voltage regulators and all seem to be operating at the correct voltage. Beyond that I need some advice on what and where to start testing. Thanks again for any help you guys can offer.
PS: Also forgot to mention there is a noticeable 60hz buzz coming from the area near the transformers. Nothing seems to be overly hot after a few minuets except the big resistors near the TO-3's.
EDIT: 2v and 120v
-
Are output and sense terminals connected properly?
-
Have you checked D17 ?
best regards
Jorn -
You should also clean up PCB around TR17 and the changed HV transistors.
You will never know what hides in the flux residue. -
Update!
In answer to your questions. Yes the terminals appear to be connected correctly. Yes I've checked most of the diodes. And yes I may have found a bad HV transistor!
TR7-11 are a mix of FQP3N90 and SSP2N90. Possibly some have been replaced at some point. I was probing around with my DMM in diode mode and saw that TR10 was conducting from gate to source in both directions! I removed (what a pain in the a$$!) and sure enough my part tester shows double diode!
https://drive.google.com/open?id=12_0FITgLfs8mv6nClZFAnrR9qFvCWycs
https://drive.google.com/open?id=1qe_wSBEApUGylwa3WPTy2g9yuhiMPNww
At this point i'm just going to replace all the HV transistors and recap all the HV caps. I was looking at FQP4N90C for a replacement. Can anyone see an issue with that?
Thanks again and I'll keep you updated.
-
Also:
Can someone recommend a part number for a replacement power switch. It works but its smashed! Having a hard time finding it.
And... When I got the unit it had a small printed label on top that said something like "High Voltage output disabled internally". I don't have it in front of me. And sure enough the 1200v selector doesn't work. Is there a way to re-enable it. I confess I haven't read the manual for this yet. If anyone knows that would be awesome. Thanks -
Update!
Got the replacement parts in and the unit seems to be in good working order.
Now that the continuous clicking from the relay is fixed I can hear a different sound. While loading the output with about 6k ohm I can hear what sounds like a high voltage discharge around the high voltage transformer. The sound increases as you select the higher ranges. Is this high voltage discharge sound normal for these units? Also the 1200v range is disabled. Can someone point me where I can re-enable it? Does anyone know if this is a hardware or firmware disable?
Here is a video describing the situation so far.
https://youtu.be/IWYDBVYSvSU -
Now that the continuous clicking from the relay is fixed I can hear a different sound. While loading the output with about 6k ohm I can hear what sounds like a high voltage discharge around the high voltage transformer. The sound increases as you select the higher ranges. Is this high voltage discharge sound normal for these units? Also the 1200v range is disabled. Can someone point me where I can re-enable it? Does anyone know if this is a hardware or firmware disable?
There is a yellow ceramic capacitor between the heatsink and the high voltage capacitors that in my unit was putting out a similar noise. Being aware there's high voltage back there, you might want to try putting pressure on it to see if it dampens the noise. If I remember correctly, this capacitor was part of a DC-DC converter circuit and I can only assume that the current ripple through the cap was doing this (or whatever mechanism causes ceramic capacitors to act like speakers)
I ended up replacing it with an identical value but different style capacitor which had a lower instance of the noise, although I probably didn't need to. -
That bodge is not present on my 2701c, see pic. Also I have shown a photo of the transformer part no which hopefully yours is the same.
One other thing is the label info on the ROM the same as the one I posted in my thread, (firmware is there too).
https://www.eevblog.com/forum/metrology/valhalla-2701c-schematics-100832/msg1380382/#msg1380382
Unfortunately mine is busted and I don't have time to repair it ATM so I cant tell if that transformer noise is normal or not.
Trace out the bodge to see what has been done first, if indeed its related to the HV range disable then get rid of it!.
edit : pics added !.
-
I think your right forrestc. C201 and C215 seem to be emitting the noise. If I nudge them with a probe the sound changes. Should I just replace them?
Thanks lowimpedance. I did find that post initially and its been helpful. I can't figure out what's going on with this bodge. I've been looking at this thing for a few hours and i'm getting tired. The 1200v switch seems good. The switch continuity appears to match the nearby switches. Also as you power up the device the led on the switch does light up although its brighter then the rest. See pic.
And I think I need help with that bodge. I'm not totally sure what it's for. It connects one side of R48 (not populated!) between TR116 and C117. And a bodge wire from TR116 to pin 5 of relay RLB. Removing both bodge circuit and bodge wire doesn't seem to effect the functionality. And still no 1200v! I've reinstalled it for now.
I've drawn a diagram of what I see on the board. See pic.
R48 shows as a 68.3k resister on the schematic but is missing on my board. See pic.
-
Are you sure of that bodge cct ?! , does not make much sense, ie c117 is in parallel with r127 and are associated with tr114 not tr116 !. Where does the base of the bodge transistor go ?, to the 0v end of r48?.
When you removed the bodge did you 'tack in' a 68K resistor for r48 as its part of the range sensing ?. Is the firmware the same as the one I posted?. And lastly your unit is not the LNF version which is limited to +/- 40V (see operators manual for very brief overview), which should be marked on the rear of the unit as such. -
First thank you for being patient with me!
The firmware is version 3.0 and the unit is 2701C-standard model. So no LNF.
Ok ignore the last bodge pic! I've drawn up a better one with the schematic overlay. So what I can see is that the RLB relay is part of this bodge which is in fact the relay responsible with the 1200v selection output.
The pic shows what's going on.
I did try removing all the bodges and installing a resister into R48. I was able to make a 68.6k instead of the required 68.3k. But there was no change!
The RLB relay is activated by IC108 which is the control unit in charge of switching the relays. Unless someone has another suggestion I'm guessing there is a change in the firmware that's doing this. I don't have the equipment to R/W the firmware chip at all. But if the 1200v can be disabled in the firmware why do the hardware bodge to neuter the relay?
I'm thinking of dropping an email to Valhalla Scientific. Its a long shot but I'm not sure what else I can do. If someone has a contact I would be grateful!
Thanks for any more help you guys can offer!
EDIT: Not in the picture is the bodge wire that is wired from RLB pin 5 to (5v).
-
Certainly give Valhalla a call to see if your serial number was a factory mod. My guess it probably was as I also suspect there is a firmware change as well as the bodge.
Is there no one you know that have a programmer that could burn a different EROM to try ?. Again Valhalla might supply a programmed EPROM.
Must have a look to see If I have a spare 27c64 floating around if all else fails.
This unit should be readily changed back to full output operation without to much effort one would think !.
Good luck -
Unfortunately I don't know anyone with such equipment. At this point I want to try a firmware fix. I was looking for a basic programmer and found this.
https://www.amazon.com/Signstek-TL866PLUS-Universal-MiniPro-Programmer/dp/B00K73TSLM/ref=sr_1_2?s=industrial&ie=UTF8&qid=1528539537&sr=1-2&keywords=TL866&dpID=41PVXsJOD4L&preST=_SX342_QL70_&dpSrc=srch
I guess its similar to a TL866 programmer from EEVBlog #411. Going to order it tonight. If the firmware you posted is from a unit with the 1200v working i'm willing to give it a shot. I have zero experience with eprom stuff. I've done a few arduino projects but this is uncharted territory for me. I've scavenged a few eproms from my junk bin that I can experiment with. I have a NM27C128Q and NM27C256N that look pin compatible. Can these be substituted for the 27C64 that's in the Valhalla?
I'll keep you updated when the programmer gets here.
-
I guess its similar to a TL866 programmer from EEVBlog #411. Going to order it tonight. If the firmware you posted is from a unit with the 1200v working i'm willing to give it a shot. I have zero experience with eprom stuff. I've done a few arduino projects but this is uncharted territory for me. I've scavenged a few eproms from my junk bin that I can experiment with. I have a NM27C128Q and NM27C256N that look pin compatible. Can these be substituted for the 27C64 that's in the Valhalla?
You can often get away with dropping a 28C64 or a 28F64 in place which are generally still available. The difference between an 27xx and 28xx is that the 28xx is electrically erasable. The 27xx needs a uv source to erase it.
I just realized I hadn't responded on the audibly noisy caps. In my case, I replaced mine just so I didn't have to listen to the noise, although it wasn't all that noticeable with the case on. I will note that the first cap that I tried did exactly the same thing. I actually think I ended up paralleling two caps which ended up being the same value as the existing one. On my board there's a second non populated spot in parallel with the first one, if I remember correctly. -
Update:
Well I received my EPROM reading/writer and have been playing with it for a few minuets and it's pretty cool. Unfortunately after reading the 2701C EPROM and comparing it with the one lowimpedance posted they appear to be identical! So i'm not even going to try and burn a new ROM.
Well at this point I'm going to send an email to Valhalla Scientific and see if they can offer some advice.
And FYI, you can erase an EPROM with a UV LED. I left it running for 24 hours and sure enough it was wiped!
-
Whats weird is that the 1200 V range LED does not come on (other than the power up reset routine), and this is only driven by the micro system via the PIA 109 then IC113. Leading to the assumption that the Firmware has some 'tweak' to it, but apparently not as the check sums are the same !. (have you tried the firmware that I posted in the unit ?.).
If the firmware's are the same then then there is either more bodges to get rid of or you still have another hardware failure, (most likely). So its safe to get rid of all bodges and make sure the unit conforms to the schematic.
Next do you have an oscilloscope?. After disabling the main HV power supply, maybe by removing the transformer leads e20,21 from the PCB and taping them to prevent oops moments, or disable the primary somehow (not looked into this however). Doing this just to be safe when probing with an Oscilloscope. So check to see if MS and LS drive and their complements are present.
Do you happen to have or able to borrow a HV probe ? to see if there is 1400 V ?. Carefully check if there is 200V across Tr6 to Tr12 ie :one side of R8 and one side of R7 ,see cct for detail. If not then maybe IC5 is duff or a possible feedback issue with IC1 ,2 and Tr102, 103 etc driving the HV transformer. Read the manual description.
Above all BE CAREFUL probing when the HV is present.
What still bugs me though is that the 1200V switch LED does not come on when the button is pressed.....
And hopefully Valhalla can shed some further light.
-
Well at this point I'm going to send an email to Valhalla Scientific and see if they can offer some advice.
I'm assuming you have full schematics... I'm wondering if by chance someone has cut the leads to (or bent outside their socket) either pin 12 or 13 on IC8 on the display board, either on the IC or on the board. If I was disabling the 1000V range, that's the way I'd do it. Or, cut a trace leading to switch16.
Because the buttons seem to be multiplexed, as long as the 20V range and the operate/standby buttons seem to work, and the firmware is identical, then that's really the only place to disable this unless there's some other configuration somewhere. It just occurred to me that the bodge might be a "make sure if it accidentally gets selected it doesn't stay selected" bodge, I'll have to look at it some more, with the benefit of the full schematic. I'll do that and re-post if I see something interesting. -
It just occurred to me that the bodge might be a "make sure if it accidentally gets selected it doesn't stay selected" bodge, I'll have to look at it some more, with the benefit of the full schematic. I'll do that and re-post if I see something interesting.
Yep...., I think I figured out the bodge.
Let's assume that they did something to the button so it couldn't be engaged. Even then, if somehow it accidentally got set, you need another way to ensure that 1000V didn't make it out of the unit.
When RLB turns on, it engages the bodge transistior. This relay is only on when 1200V is set. Once this transistor is on, it forces the D input to the flip-flop IN IC112 low.
If you read section 9-3-5 of the manual, you'll find that IC112 and the surrounding circuit is the DAC for the unit. By yanking this line low, you effectively force MS_DRIVE low, and ~MS_DRIVE high. Switching to the main pcb schematic, you find that the MS_DRIVE signal pairs toggle a pair of fets between the 6.95vref and 0V. This bodge thus forces this pair of fets to always be at 0V, resulting in the signal at the right hand side of R35 to never be very high (even if the high side of LS_DRIVE gets set on always). This signal is then used as the reference for the feedback circuit around IC5. If this signal is close to zero, the voltage output will also be close to zero.
So effectively, this bodge neuters the output so when RLB is on, the reference voltage drops to close to zero, making sure the output never gets very high, even if 1000V accidentally gets selected.
It would be interesting to see if you find a hack on the display board as I described, and if you fix it, you leave the bodge in place (exactly as you found it) just long enough to see if this does in fact limit the voltage, and if so, how much it limits it. I suspect you'll get almost zero volts out with this bodge in place, even when set to 1000V.
Oh, and when you do remove the bodge, make sure the +5V wire has been removed, and make sure all of the connections are back the way they should be. It wouldn't surprise me if they had cut the trace on that relay pin that the +5V wire is attached to. -
OK I FOUND SOMETHING!
Props to you forrestc for figuring out what was most likely wrong at nearly the SAME TIME I was at the bench finding what was wrong!
So I should have started there first. I was too preoccupied with the main board and replacing the HV transistor section that I never looked under the face plate for other issues. Well I indeed found what was keeping the 1200v switch from working! A damaged/cut trace at the 1200v selection switch!!!!! And it almost fixed it!!!
Sure enough its a trace connecting the switch to pin 13 on IC8! I quickly wired a bodge wire and reassembled it. Well the HV selection activates but the output is not what it should be. The output voltages are roughly 10X less than displayed. The 120v selection matches the output but with the 1200v selected the output is far reduced even when adjusting to a voltage near the 120v range. So now I need to figure this out. I thought it could be something in the calibration settings. I quickly followed the calibration procedure but can't get the 1200v range to adjust.
We're definitely making progress here and thanks for helping out guys. In answer to the previous questions I do have a scope and an older HV probe. I will probe around and see if anything stands out. Thanks lowimpedance i'll try testing around the HV transistors again but after I replaced them I did test. There is crazy HV all over this board! But I don't remember getting anything over 1000v. With the unit powered up and the 1200v NOT selected is there a place on the board that should have in excess of 1000v? I think I was getting something like 750v across the two big orange resistors on the left side of the main board. I'll try to get back to it tomorrow night.
I've worked with tube amplifiers and I try to be very careful around HV DC stuff like this but thanks for the warnings. Complacency would definitely lead to trouble or worse with this stuff! I must admit that I'm still a bit new at trouble shooting something this complicated. And I'm still very appreciated for any farther help.
-
Sure enough its a trace connecting the switch to pin 13 on IC8! I quickly wired a bodge wire and reassembled it. Well the HV selection activates but the output is not what it should be. The output voltages are roughly 10X less than displayed. The 120v selection matches the output but with the 1200v selected the output is far reduced even when adjusting to a voltage near the 120v range. So now I need to figure this out. I thought it could be something in the calibration settings. I quickly followed the calibration procedure but can't get the 1200v range to adjust.
Please see my second response about the original bodge wire/circuit you found. The original bodge you found is specifically designed to severely limit the output in the 1000V range. I described it in more detail in that post, but in short, the transistor in the bodge gets turned on by the relay, and that transistor in turn prevents the Digital to Analog converter circuitry from outputting a voltage high enough to get very much voltage out of the unit. (The HV circuitry effectively "amplifies" the voltage out of the DAC).
If you remove the original bodge you found, including that 5V jumper on the relay, and make sure that you've replaced the resistor they removed, then 1000V should work for you. Be sure to make sure there isn't anything else cut around RLB as well. Really, what you're trying to do is put the affected portions of the circuit back the way they are on the schematic.
I would be surprised if you don't get a working unit as a result. -
Thanks forrestc. I should have been a little more scrutinous. I definitely lack some knowledge about how this thing actually works. I did find another cut trace and pin 5 of RLB. I just wasn't looking hard enough!
So I reconnected that and powered up. SPARKS!!! Well just one but I definitely flinched! 
I don't remember hitting it but C12 seems to have made contact with an iron. It could have been me but this was a salvaged unit so who knows. I've patched it with some silicon adhesive for now but it will have to be replaced of course. I shop on Digikey and I'm having an issue finding a suitable replacement. Anyone have a suggestion?
I can read a schematic but i'm trying to expand my knowledge of the "how and why" of the engineering. So let me see if I can talk through this.
Starting with the high voltage source at capacitor stack C203-C206 feeds into the high voltage transistors TR7-TR11. TR12 and TR6 create a Darlington correct? This seems to be the control point for the main voltage source. This is controlled from IC5 (ICL7650S) Chopper-Stabilized Operational Amplifier. This amp is being controlled by TR13-TR16 which in turn is controlled by the DAC. The main feedback line is tapped from D8 (on the schematic). The voltage selections switch on and off RLA, RLB, and RLC that provide feedback for the amp. Now the defeat bodge has pin 5 and 11 from RLB cut and linked to +5v. Pins 4 and 10 are linked to the base of a transistor inserted at pin 5 of IC112 and ground. So if the relay was triggered by any means the +5v would trigger the transistor pulling pin 5 of IC112 to ground. This also means that RLB was fully removed from the feedback path. So this bodge did not introduce anything into the feedback path and RLB was isolated. So they introduced there own feedback mechanistic using RLB's functionality so in the event it was forced on it would prevent the control hardware from generating a signal that would result in a high voltage output (pin 5 on IC112). I think that's right or someone can correct me.
So in summary the unit seems fully functional. In the picture i'm using a 1000:1 probe. The initial problem was a continuous reset of RLJ possibly due to a dead FET. New HV transistors fixed that. Recapped the electrolytics. Removed two defeat bodges and repaired two cut traces. Replaced R48 with something very close to the schismatic. Replaced C201 and C215 due to the noise(didn't fully remove the noise but much reduced). And now I have a cool EPROM reader.
I'll finish with a question. With these repairs done is it worth having it calibrated? WOULD anyone calibrate this unit with these repairs? Would they look inside or just do the manual calibration? What would that cost? I bought this because is was not functioning and thought it would be fun to fix it up. I'm not sure if this thing is still in demand somewhere or if it's simply too old for real use. Now that it's working I'm not totally sure what I should do with it. I have a few HV power supplies. I did learn something though. Thanks for helping out guys!
-
I looked at the original picture and the breach on C12 was there before I did my repair! The HV transistors did have a few different manufactures so someone else had tried to repair but possibly damaged C12 and thus arced and damaged TR10? Or with this high voltage could the breach in C12 have happened over time?
-
Good job on the repair, I'm sure you have learned something from it and this thread will have useful info for others (including me when I finally get round to fixing mine).
Looking at the pic. of C12 it looks like heat damage from a soldering iron and not an age thing. Replace asap.
I think this unit is still sold by Valhalla looking at their web site , request a quote only
. So probably still worth something now its functional.
One note that R48 should have similar TCR and tolerance to the original part spec. if possible. -
Thanks forrestc. I should have been a little more scrutinous. I definitely lack some knowledge about how this thing actually works. I did find another cut trace and pin 5 of RLB. I just wasn't looking hard enough!
So I reconnected that and powered up. SPARKS!!! Well just one but I definitely flinched! I don't remember hitting it but C12 seems to have made contact with an iron. It could have been me but this was a salvaged unit so who knows. I've patched it with some silicon adhesive for now but it will have to be replaced of course. I shop on Digikey and I'm having an issue finding a suitable replacement. Anyone have a suggestion?
Start a new thread with a picture of the cap (maybe in a different category), and ask if anyone has a source. I checked around through a few of my likely sources, couldn't find anything.Starting with the high voltage source at capacitor stack C203-C206 feeds into the high voltage transistors TR7-TR11. TR12 and TR6 create a Darlington correct? This seems to be the control point for the main voltage source. This is controlled from IC5 (ICL7650S) Chopper-Stabilized Operational Amplifier. This amp is being controlled by TR13-TR16 which in turn is controlled by the DAC. The main feedback line is tapped from D8 (on the schematic). The voltage selections switch on and off RLA, RLB, and RLC that provide feedback for the amp. Now the defeat bodge has pin 5 and 11 from RLB cut and linked to +5v. Pins 4 and 10 are linked to the base of a transistor inserted at pin 5 of IC112 and ground. So if the relay was triggered by any means the +5v would trigger the transistor pulling pin 5 of IC112 to ground. This also means that RLB was fully removed from the feedback path. So this bodge did not introduce anything into the feedback path and RLB was isolated. So they introduced there own feedback mechanistic using RLB's functionality so in the event it was forced on it would prevent the control hardware from generating a signal that would result in a high voltage output (pin 5 on IC112). I think that's right or someone can correct me.
I'm away from the schematic right now so I can't verify your details, but this sounds pretty close to reality based on my recollection of the entire circuit.I'll finish with a question. With these repairs done is it worth having it calibrated? WOULD anyone calibrate this unit with these repairs? Would they look inside or just do the manual calibration? What would that cost? I bought this because is was not functioning and thought it would be fun to fix it up. I'm not sure if this thing is still in demand somewhere or if it's simply too old for real use. Now that it's working I'm not totally sure what I should do with it. I have a few HV power supplies. I did learn something though.
I use mine as a 'transfer standard'. I have a keithley 2001 I send off to get calibrated every year. As soon as it comes back from the cal lab, I do some tests to verify it is in within spec, and then touch up the calibration on my 2701C using it. I then use the 2701C to verify/calibrate everything else in the lab which can be calibrated with a voltage source. When I'm done calibrating everything, I then re-test the 2701c against the 2001, to make sure it didn't drift during the calibration.
The calibration procedure for the 2701c only needs a high accuracy DMM, unless you've got the current option, and care about it (I don't care, so I skip that section). You basically switch it to cal mode, then following the procedure in the manual, you just adjust the knobs on the 2701C such that the displayed voltage on the DMM matches the expected voltage out of the 2701C.
Between calibration cycles, I use the 2701c for a lab high precision voltage source.
I don't know what a cal lab would charge on the 2701c, but the 2001 costs around $150 depending on the lab I use.