Where are you taking your ground reference point from when doing this measurement?
Try to find ground point on the board nearest to the U100.
Also repeat you measurements at tp101.
With the input shorted to ground you should see no significant dc at these test points.
The op amp U100, unless you have replaced it, has a darlington pair at its input.
The input bias current is very low and shouldn't produce much dc offset even at a worst case input R.
The key to trouble shoot this is to try and isolate this circuit and see if the problem still exists.
Is it at the input?, is it at the output? sort of thing of both the passive and active stages.
You may have a potential difference of grounding points.
Also check if the problem is with the source signal, the oscillator. Is there a dc component in the signal?
Is it a difference between grounds?
The 339A has two isolated grounds. The oscillator ground floats.
Try to find ground point on the board nearest to the U100.
Also repeat you measurements at tp101.
With the input shorted to ground you should see no significant dc at these test points.
The op amp U100, unless you have replaced it, has a darlington pair at its input.
The input bias current is very low and shouldn't produce much dc offset even at a worst case input R.
The key to trouble shoot this is to try and isolate this circuit and see if the problem still exists.
Is it at the input?, is it at the output? sort of thing of both the passive and active stages.
You may have a potential difference of grounding points.
Also check if the problem is with the source signal, the oscillator. Is there a dc component in the signal?
Is it a difference between grounds?
The 339A has two isolated grounds. The oscillator ground floats.
Last edited:
David,
I was taking my ground from the shield attached to the ground plane near S4. I'll try getting my ground right on or near U100 for the measurement.
On your comment about shorting the input to ground to see if the DC remains. I was contemplating removing the input signal to A4, the J100 connector, and shorting that RCA out.
I had family unexpectedly come by this evening so no time to do any tests. I'll get more info tomorrow. Thanks for your input!
I was taking my ground from the shield attached to the ground plane near S4. I'll try getting my ground right on or near U100 for the measurement.
On your comment about shorting the input to ground to see if the DC remains. I was contemplating removing the input signal to A4, the J100 connector, and shorting that RCA out.
I had family unexpectedly come by this evening so no time to do any tests. I'll get more info tomorrow. Thanks for your input!
I had some time tonight to run a few more checks. I pulled the J100 signal input and installed an RCA shorting plug. These measurements are with CR102 & CR103 still lifted and out of circuit. Rotating the input range selector down starting at 300V the neg DC rises at TP 101 with each step until a peak of -250mV DC on the 10V input range after switching below this range it drops to 0V at all ranges below 10V. TP103 on the 10V range is -400mV.
TP103 shows some amount of DC offset at the lower ranges below 10V (although < 400mV) all except the very lowest range which is the 100uV range on my Opt 1 339. The 100uV range is 0v of DC on TP 103. This lowest range step appears to be a 50.51 ohm resistor with no parallel capacitance on S4E.
TP103 shows some amount of DC offset at the lower ranges below 10V (although < 400mV) all except the very lowest range which is the 100uV range on my Opt 1 339. The 100uV range is 0v of DC on TP 103. This lowest range step appears to be a 50.51 ohm resistor with no parallel capacitance on S4E.
Did put a socket in place for u100?
I seem to remember Sync having a similar problem with a different stage. It turned out there was a small solder bridge under the socket he placed which he couldn't see for the socket.
Maybe remove the op amp and clean the area. Look for debris or bridges. Check both sides of the board. Look for cracked trace as well near the work.
The op amp is an HA-2625. an old Harris part. Input bias current is 5nV. If the input R were 1M ohm the offset would be 5mV. You shouldn't see anything more than that or less.
Could be a floating ground. See if you get ground continuity right back to the power supply entry.
Lift one leg of F100 or R110. If you still have voltage greater than 5mV or so then it's being feed back from the op amp circuit. Something is partially shorting there.
What's the voltage at the power pins of the op amp?
I seem to remember Sync having a similar problem with a different stage. It turned out there was a small solder bridge under the socket he placed which he couldn't see for the socket.
Maybe remove the op amp and clean the area. Look for debris or bridges. Check both sides of the board. Look for cracked trace as well near the work.
The op amp is an HA-2625. an old Harris part. Input bias current is 5nV. If the input R were 1M ohm the offset would be 5mV. You shouldn't see anything more than that or less.
Could be a floating ground. See if you get ground continuity right back to the power supply entry.
Lift one leg of F100 or R110. If you still have voltage greater than 5mV or so then it's being feed back from the op amp circuit. Something is partially shorting there.
What's the voltage at the power pins of the op amp?
Last edited:
I pulled the DIP socket I installed for U100 and verified there were no solder bridges or cracked traces. I reinstalled it and ohmed it from the socket side of the pins back to the first component to make sure they all had good continuity.
Pulling F100 and measuring TP 101 shows much higher neg voltage, several volts actually depending on the range setting. I checked ground continuity back to the power supply and it measures good, < 1 ohm anyway.
The voltage at the pins of the op-amp is -15.25V & +15.31V.
I tried reinstalling the LME49710 into U100 position see what would happen and the measured results were similar. I tried with both pin 8 connected and with it bent and out of circuit.
Pulling F100 and measuring TP 101 shows much higher neg voltage, several volts actually depending on the range setting. I checked ground continuity back to the power supply and it measures good, < 1 ohm anyway.
The voltage at the pins of the op-amp is -15.25V & +15.31V.
I tried reinstalling the LME49710 into U100 position see what would happen and the measured results were similar. I tried with both pin 8 connected and with it bent and out of circuit.
I pulled the DIP socket I installed for U100 and verified there were no solder bridges or cracked
With F100 lifted there is nothing but your probe to supply bias current to the op amp input. This will account for a large amount of offset at the output of the op amp but not at the input. With the op amp removed do you still have voltage at TP101?
The circuit connected to pin 8 is just a voltage clamp. It connects to the op amp's compensation pin which is internally connected to the amplifiers voltage stage. This stage is trans-conductance in nature so high impedance. Pin 8 is not the same purpose for all op amps.
I would pull the op amp and check to see if voltage is being feed back from a later stage on the output pin. Check all the pads for a source of voltage. There should be none other than the power pin pads and the voltage clamp. In fact it might be a problem with the clamp circuit. See if the clamp circuit voltages make sense for the component values. the circuit is supplied from the negative rail.
Sometimes when we remove ICs from a double sided board the through hole plating can get cracked or the pad can separate from the through hole plating. Often this is not visible to the eye. To be sure check the continuity from one side of the board to the other for each pad.
The input range switch has a wafer which changes the ac/dc gain of the U100. Not every switch position changes the gain but it will change at certain positions. This accounts for the offset change at the output of U100. The input bias current offset is amplified. So as the amplifier gain changes so will the output offset. This is normal.
Try this: with the op amp installed solder a 1k resistor from ground to pin 3, the non inverting input. Leave F100 out and all other previously lifted components lifted. We want the amplifier in an isolated state. See if everything returns to normal with this 1 k resistor installed.
With F100 lifted there is nothing but your probe to supply bias current to the op amp input. This will account for a large amount of offset at the output of the op amp but not at the input. With the op amp removed do you still have voltage at TP101?
The circuit connected to pin 8 is just a voltage clamp. It connects to the op amp's compensation pin which is internally connected to the amplifiers voltage stage. This stage is trans-conductance in nature so high impedance. Pin 8 is not the same purpose for all op amps.
I would pull the op amp and check to see if voltage is being feed back from a later stage on the output pin. Check all the pads for a source of voltage. There should be none other than the power pin pads and the voltage clamp. In fact it might be a problem with the clamp circuit. See if the clamp circuit voltages make sense for the component values. the circuit is supplied from the negative rail.
Sometimes when we remove ICs from a double sided board the through hole plating can get cracked or the pad can separate from the through hole plating. Often this is not visible to the eye. To be sure check the continuity from one side of the board to the other for each pad.
The input range switch has a wafer which changes the ac/dc gain of the U100. Not every switch position changes the gain but it will change at certain positions. This accounts for the offset change at the output of U100. The input bias current offset is amplified. So as the amplifier gain changes so will the output offset. This is normal.
Try this: with the op amp installed solder a 1k resistor from ground to pin 3, the non inverting input. Leave F100 out and all other previously lifted components lifted. We want the amplifier in an isolated state. See if everything returns to normal with this 1 k resistor installed.
Last edited:
Hi David,
So I started on your last list of items to check. I had previously reconnected the CR102 & CR103 protection diodes so I pulled U100 and F100 and started checking DC voltages.
Pin 3, -7.7V
Pin 4, -15.35V
Pin 5, 0V
Pin 6, 73mV
Pin 7, 15.25V
Pin 8, 9.56V
Pin 3 is extremely high so I lifted the CR102/103 diodes again to get them out of circuit. Pin 3 still measures -7.7VDC. My thinking was that there has to be something wrong with a shorted solder trace or something so I inspect the hell out of the board on both sides. I then realized that what I'm seeing on the board didn't match up with the schematic I was following. After about a half hour of following traces around the board it dawned on me that I was chasing my tail. The 339A Opt 1 schematic has more changes than I had realized. See the attached schematic for the Opt 1 A4 board.
In the Opt 1 339a the dual gate FET in the protection circuit is directly tied to Pin 3 and NOT Pin 2 as in the non-Opt 1 instruments. I didn't pull the Dual gate FET but I'm sure that's where the neg voltage is coming from, its the only thing tied to pin 3/TP 101 now. I will lift it tomorrow to verify this though.
So I started on your last list of items to check. I had previously reconnected the CR102 & CR103 protection diodes so I pulled U100 and F100 and started checking DC voltages.
Pin 3, -7.7V
Pin 4, -15.35V
Pin 5, 0V
Pin 6, 73mV
Pin 7, 15.25V
Pin 8, 9.56V
Pin 3 is extremely high so I lifted the CR102/103 diodes again to get them out of circuit. Pin 3 still measures -7.7VDC. My thinking was that there has to be something wrong with a shorted solder trace or something so I inspect the hell out of the board on both sides. I then realized that what I'm seeing on the board didn't match up with the schematic I was following. After about a half hour of following traces around the board it dawned on me that I was chasing my tail. The 339A Opt 1 schematic has more changes than I had realized. See the attached schematic for the Opt 1 A4 board.
In the Opt 1 339a the dual gate FET in the protection circuit is directly tied to Pin 3 and NOT Pin 2 as in the non-Opt 1 instruments. I didn't pull the Dual gate FET but I'm sure that's where the neg voltage is coming from, its the only thing tied to pin 3/TP 101 now. I will lift it tomorrow to verify this though.
Attachments
There's other ways of replacing it. Just takes a bit of work.
An op amp can be used for same purpose. it just cost more.
@Dave.
It copies the input signal in phase and feeds a portion of the signal to diodes. Since the diodes have virtually the same signal across them, the leakage is null and doesn't cause distortion to the input signal.
The Z can't be too low as to counteract the operation of the limiting circuit. This is handled by the 300pF coupling caps.
The intention of the circuit is to fulfill a safety requirement of the uint being able to withstand a 300V dc/ac input without taking the operator out.
Otherwise what would it do besides frying the input op amp and maybe a fire.
It copies the input signal in phase and feeds a portion of the signal to diodes. Since the diodes have virtually the same signal across them, the leakage is null and doesn't cause distortion to the input signal.
The Z can't be too low as to counteract the operation of the limiting circuit. This is handled by the 300pF coupling caps.
The intention of the circuit is to fulfill a safety requirement of the uint being able to withstand a 300V dc/ac input without taking the operator out.
Otherwise what would it do besides frying the input op amp and maybe a fire.
Last edited:
Okay the problem is solved.
I removed the Q100 dual gate FET and I still had -7V at pin 3 of the U100 socket (U100 removed). I removed the U100 socket again (3rd time now) and inspected, all looked fine. I powered up the 339 and checked the U100 pin 3 location on the bare board, no DIP socket, and measured 0 volts. I ohmed the DIP socket I removed out between Pin 3 & 4 and I measure about 200k ohm with my DMM. All other pins measure open on the socket to their adjacent pins. I can see no visible link between Pin 3 & 4, but its surely conductive between them.
I grabbed a new DIP socket out of the parts bin and it measured open between Pin 3 & 4. Installed it into the U100 location and powered up the 339 with no U100 opamp and there's now no DC voltage at Pin 3. I reinstalled an LME49710 into U100 and the 339 is now working perfectly on all ranges. No DC offset whatsoever, the sine wave stays nice and centered on the scope no matter which range I choose. Its also measuring distortion on all ranges again.
I still need to put the protection circuit back together, hopefully the dual gate mosfet is okay. It only saw about -7V so I think it should be fine.
David I appreciate your help and the good information!
I removed the Q100 dual gate FET and I still had -7V at pin 3 of the U100 socket (U100 removed). I removed the U100 socket again (3rd time now) and inspected, all looked fine. I powered up the 339 and checked the U100 pin 3 location on the bare board, no DIP socket, and measured 0 volts. I ohmed the DIP socket I removed out between Pin 3 & 4 and I measure about 200k ohm with my DMM. All other pins measure open on the socket to their adjacent pins. I can see no visible link between Pin 3 & 4, but its surely conductive between them.
I grabbed a new DIP socket out of the parts bin and it measured open between Pin 3 & 4. Installed it into the U100 location and powered up the 339 with no U100 opamp and there's now no DC voltage at Pin 3. I reinstalled an LME49710 into U100 and the 339 is now working perfectly on all ranges. No DC offset whatsoever, the sine wave stays nice and centered on the scope no matter which range I choose. Its also measuring distortion on all ranges again.
I still need to put the protection circuit back together, hopefully the dual gate mosfet is okay. It only saw about -7V so I think it should be fine.
David I appreciate your help and the good information!
I got the 339A protection circuit reassembled and working again this morning. Its working great now, showing .0007% THD+N with a 1kHz signal. Its also very stable with the needle occasionally wavering about +/- .0001%.
I fed the 339 oscillator into my Amber 3501 and it shows about -104.5dB. My Amber's internal oscillator only measures about -102.5dB in loopback. So it appears that my 3501 distortion analyzer outperforms its own internal oscillator and inversely my 339 internal oscillator outperforms the 339a distortion analyzer. Oh well they're both working good, so its hard to complain.
I fed the 339 oscillator into my Amber 3501 and it shows about -104.5dB. My Amber's internal oscillator only measures about -102.5dB in loopback. So it appears that my 3501 distortion analyzer outperforms its own internal oscillator and inversely my 339 internal oscillator outperforms the 339a distortion analyzer. Oh well they're both working good, so its hard to complain.