- Joined
- Nov 28, 2011
- Messages
- 8,393
You've got Q8 and Q6 swapped over. Other than that it looks OK.
What book do I need to buy to be able to figure out which transistor is being used? I look up the HP part number... all I get is vendors that want me to submit a request for quote. I look it up on the HP cross-reference I downloaded... not there. I search for the Motorola part number... no hits. I look for the Motorola book that described their transistors... one on Archive.Org, but the others are on sites that require subscription or special procedures that drive me nuts.
I hate to keep asking for cross reference information, so how can I become independent in that regard?
I hate to keep asking for cross reference information, so how can I become independent in that regard?
I found a place in Texas that stocks a lot of the old HP parts. Lucky for me, the ones I could not translate into modern equivalents were available, so I bit the bullet and ordered them. The rest seem to translate into 2N3904 and 2N3906, for the most part. So I think I'm ready to move forward once those parts arrive. The downside, of course, is that ordering parts this way is expensive and time consuming.
Today was more about prep than progress. I mapped out all the remaining transistors on A15 (HVPS) and A16 (LVPS) that might need replacing, and lined up the suppliers. Regarding the LVPS 48V line, I got all but one of the transistors sourced, and now just have to wait for them to arrive. I do have a replacement for Q7, but not Q5-6 or Q8-9. I was only able to get one of the 1853-0080s (Q5-6), so I'll flip a coin on that one.
I decided to replace A16Q11 in the 156V line just because I got such poor readings on that line, and because the part (a 2N3440 equivalent) was available at Vetco for $3.35
I decided to replace A16Q11 in the 156V line just because I got such poor readings on that line, and because the part (a 2N3440 equivalent) was available at Vetco for $3.35
- Joined
- Jan 21, 2010
- Messages
- 25,510
It might be useful if you can tell us who that supplier is. It may be useful for someone else needing to repair HP equipment.
It would also be really interesting to know if the parts you get are NOS HP parts with HP part numbers on them, or normal OEM parts with conventional markings. (edit: or even new HP numbered parts)
It would also be really interesting to know if the parts you get are NOS HP parts with HP part numbers on them, or normal OEM parts with conventional markings. (edit: or even new HP numbered parts)
The dealer is Tucker Electronics at etestparts.com. I'll let you know whether they are NOS (in original bags) or otherwise once they arrive. The HP parts are listed under the Agilent/HP category.
I've seen at least one vendor online that does mention on their web site when parts are new in bags... but they have very few parts.
I've seen at least one vendor online that does mention on their web site when parts are new in bags... but they have very few parts.
- Joined
- Nov 28, 2011
- Messages
- 8,393
If you're going to replace A15U1 (CA3094A) it would be worth checking the voltages around it, and the operation of the HVPS without it, before you install the new IC. This is just a precaution because the CA3094A is a scarce commodity.
I also suggest that you fix the +48V regulator before moving on to the HVPS.
Once you've got the +48V rail up and running, here are a few ideas for how to check the HVPS without the CA3094A present.
With the old CA3094A removed, and the scope powered up, the HVPS should not run - that is, no glowing neons and no ozone smell.
In this condition you can measure the voltages on the U1 pads. Measure the voltages with the multimeter negative lead connected to the circuit's 0V rail, the chassis symbol that connects to several components near U1: R24, R20, and various small capacitors. The voltages should be:
Pin 3: +14V to +16V
Pin 4: -14V to -16V
Pin 7: +14V to +16V
Pin 8: +14V to +16V
The other pins aren't important.
Next, you can try running up the converter without regulation. You'll need a 20k multi-turn trimpot such as http://www.digikey.com/product-detail/en/64WR20KLF/987-1064-ND/2408642. You might also want to get a trimpot adjusting tool http://www.digikey.com/product-detail/en/H-90/H90-ND/262571 which won't slip off the adjusting head - it makes it much easier to adjust a multi-turn trimpot.
Turn the trimpot fully anticlockwise until it starts to click. Measure resistance between the middle wire and the two end wires. One will read around 20 kilohms, and the other will read close to zero ohms. Bend the latter one away. Connect the other two wires into pins 8 and 6 of the CA3094A position.
Connect your multimeter on voltage range with the negative probe to the 0V rail (as before), and the positive probe to the junction of R23 and R37.
Leave the HV MULT wire disconnected.
Power up the scope and note the voltage on the multimeter. It should be between +10V and +16V.
Now gradually turn the trimpot clockwise. This will increase the bias to the HV OSC transistor and increase the energy going into the HV transformer, which will cause the HV output voltages to increase. Unfortunately, there's nowhere to safely measure the HV output voltage because it's too high, but the voltage shown on the multimeter should drop as you turn the trimpot clockwise.
Keep turning the trimpot clockwise until either you notice an ozone smell, or any neons start to flash or glow, or the voltage reaches 0V. If the voltage reaches 0V and remains fairly steady, the rest of the HVPS is probably OK and it's safe to install the CA3094A. If you notice ozone or neons, make a note of the voltage on the multimeter and post it here.
I also suggest that you fix the +48V regulator before moving on to the HVPS.
Once you've got the +48V rail up and running, here are a few ideas for how to check the HVPS without the CA3094A present.
With the old CA3094A removed, and the scope powered up, the HVPS should not run - that is, no glowing neons and no ozone smell.
In this condition you can measure the voltages on the U1 pads. Measure the voltages with the multimeter negative lead connected to the circuit's 0V rail, the chassis symbol that connects to several components near U1: R24, R20, and various small capacitors. The voltages should be:
Pin 3: +14V to +16V
Pin 4: -14V to -16V
Pin 7: +14V to +16V
Pin 8: +14V to +16V
The other pins aren't important.
Next, you can try running up the converter without regulation. You'll need a 20k multi-turn trimpot such as http://www.digikey.com/product-detail/en/64WR20KLF/987-1064-ND/2408642. You might also want to get a trimpot adjusting tool http://www.digikey.com/product-detail/en/H-90/H90-ND/262571 which won't slip off the adjusting head - it makes it much easier to adjust a multi-turn trimpot.
Turn the trimpot fully anticlockwise until it starts to click. Measure resistance between the middle wire and the two end wires. One will read around 20 kilohms, and the other will read close to zero ohms. Bend the latter one away. Connect the other two wires into pins 8 and 6 of the CA3094A position.
Connect your multimeter on voltage range with the negative probe to the 0V rail (as before), and the positive probe to the junction of R23 and R37.
Leave the HV MULT wire disconnected.
Power up the scope and note the voltage on the multimeter. It should be between +10V and +16V.
Now gradually turn the trimpot clockwise. This will increase the bias to the HV OSC transistor and increase the energy going into the HV transformer, which will cause the HV output voltages to increase. Unfortunately, there's nowhere to safely measure the HV output voltage because it's too high, but the voltage shown on the multimeter should drop as you turn the trimpot clockwise.
Keep turning the trimpot clockwise until either you notice an ozone smell, or any neons start to flash or glow, or the voltage reaches 0V. If the voltage reaches 0V and remains fairly steady, the rest of the HVPS is probably OK and it's safe to install the CA3094A. If you notice ozone or neons, make a note of the voltage on the multimeter and post it here.
Totally agree about fixing the 48V rail first. In fact, I kinda had my head set on getting all of the rails of the LVPS fixed first, if possible. Like you said, I don't want to waste the replacement CA3094AT.
I already have a bag full of trim pots. I'll see if any of them are the right kind, but I'm pretty sure I have several of that range. It will be a while before I can run those tests, though, since I'm waiting for the parts to fix the LVPS.
I already have a bag full of trim pots. I'll see if any of them are the right kind, but I'm pretty sure I have several of that range. It will be a while before I can run those tests, though, since I'm waiting for the parts to fix the LVPS.
I did find a suitable trimmer in the bag, so at least that is one thing I don't need to acquire.
I checked on the order for the CA3094AT. The seller hasn't shipped yet, and won't until next week at the earliest. Not expected to get here until some time in January. That's frustrating.
The remaining parts for the LVPS 48V rail repair should get here tomorrow or Monday. Since I do have a 2N3904, I'll go ahead and swap out A16Q7 and then start looking at the +15V UN GND line.
I checked on the order for the CA3094AT. The seller hasn't shipped yet, and won't until next week at the earliest. Not expected to get here until some time in January. That's frustrating.
The remaining parts for the LVPS 48V rail repair should get here tomorrow or Monday. Since I do have a 2N3904, I'll go ahead and swap out A16Q7 and then start looking at the +15V UN GND line.
We are definitely making progress on the LVPS. I just did a fresh batch of readings with all the secondaries removed. Still some discrepancies, but much closer.
At the +156 rail I now get +157
At the +120 rail I get +133.6
At the +15 rail it reads +14.9
At the -15 rail it reads -14.8
At the -100 test point I get -95.6
At the +15 UNREG rail it still reads +27.4
At the +48 rail we nailed it at 48V.
Looks like those parts I ordered for the 48V rail won't be needed. The good news is we don't have to wait for them to move forward.
I would like to know what's causing the discrepancies at the +120 and +15 UNREG rails.
At the +156 rail I now get +157
At the +120 rail I get +133.6
At the +15 rail it reads +14.9
At the -15 rail it reads -14.8
At the -100 test point I get -95.6
At the +15 UNREG rail it still reads +27.4
At the +48 rail we nailed it at 48V.
Looks like those parts I ordered for the 48V rail won't be needed. The good news is we don't have to wait for them to move forward.
I would like to know what's causing the discrepancies at the +120 and +15 UNREG rails.
- Joined
- Nov 28, 2011
- Messages
- 8,393
The +120V rail used to be correct, didn't it? If you've disconnected all the outputs from other parts of the scope, that might be why the +120V rail has increased. You really need to measure them under load.
The +156V rail is generated by piggy-backing on top of the +120V rail. So if the +120V rail was correct, not 13.6V too high, the +156V rail would drop to 143.4V. So there is probably still a problem there. You need to connect the rest of the scope to know for sure.
The "+15 UNREG" rail is not actually supposed to be +15V. The name they have used is misleading. It's actually the voltage that feeds into the +15V regulator. It should be at least +20V. So there is no problem there.
The +156V rail is generated by piggy-backing on top of the +120V rail. So if the +120V rail was correct, not 13.6V too high, the +156V rail would drop to 143.4V. So there is probably still a problem there. You need to connect the rest of the scope to know for sure.
The "+15 UNREG" rail is not actually supposed to be +15V. The name they have used is misleading. It's actually the voltage that feeds into the +15V regulator. It should be at least +20V. So there is no problem there.
- Joined
- Jan 21, 2010
- Messages
- 25,510
The +15 unreg rail *MAY* be a rail designed to supply other 15V regulators.
The 120V rail may be drifting high due to no load? Is the power supply disconnected at the moment when you're taking these readings?
Sorry, I'm not following along as closely as I should if I'm going to jump in and offer an opinion.
edit: too late
The 120V rail may be drifting high due to no load? Is the power supply disconnected at the moment when you're taking these readings?
Sorry, I'm not following along as closely as I should if I'm going to jump in and offer an opinion.
edit: too late
I reassembled the scope to test it under load (except for HV MULT, which remains disconnected). Your predictions rang true. Readings at the rail are as follows:
+156 rail has +144
+120 rail shows +120.1
+48 rail comes in at +48.1
+15 Unreg sill says 24.7, as you expected
On the other side of the board the -100 test point dropped to 1.1, and the -15 and -5 showed similar readings.
At no time during this test (which took a few minutes) did I detect any ozone smell. The only neon that came on was the one on the riser board between LVPS and HVPS. The 5 neons on HVPS remained dark and silent.
And yes, I made sure the HV OSC was plugged in.
+156 rail has +144
+120 rail shows +120.1
+48 rail comes in at +48.1
+15 Unreg sill says 24.7, as you expected
On the other side of the board the -100 test point dropped to 1.1, and the -15 and -5 showed similar readings.
At no time during this test (which took a few minutes) did I detect any ozone smell. The only neon that came on was the one on the riser board between LVPS and HVPS. The 5 neons on HVPS remained dark and silent.
And yes, I made sure the HV OSC was plugged in.
- Joined
- Nov 28, 2011
- Messages
- 8,393
OK, for the 156V rail, can you do the following.
1. Add a 1N4007 diode permanently to the LVPS board, with its anode to the +120V rail (on one end of CR3) and its cathode to the emitter of Q11 (the +156V rail). This will give some protection to Q11.
2. With power on, measure the voltage across C7. It should be at least 45V.
3. With power on, measure the voltage across VR3.
4. Power off and remove Q11. Check the resistances of R18 and R17 are correct.
5. Power on (with Q11 removed) and measure voltage across VR3 again.
It's odd that the HVPS is not doing anything now. Have you disconnected or removed anything from it? If not, you should - running it when it's not regulating properly will stress the transformer and other components.
1. Add a 1N4007 diode permanently to the LVPS board, with its anode to the +120V rail (on one end of CR3) and its cathode to the emitter of Q11 (the +156V rail). This will give some protection to Q11.
2. With power on, measure the voltage across C7. It should be at least 45V.
3. With power on, measure the voltage across VR3.
4. Power off and remove Q11. Check the resistances of R18 and R17 are correct.
5. Power on (with Q11 removed) and measure voltage across VR3 again.
It's odd that the HVPS is not doing anything now. Have you disconnected or removed anything from it? If not, you should - running it when it's not regulating properly will stress the transformer and other components.
After checking the HVPS setup again I found that one of the wires connecting the HV OSC had broken. However, when I soldered it back in place again there was no change.
I haven't had a chance to make the changes/tests you recommended to the LVPS. It's getting late here and I find I'm more prone to making mistakes in the evening. I'll come back to it tomorrow with fresher eyes and mind.
I haven't had a chance to make the changes/tests you recommended to the LVPS. It's getting late here and I find I'm more prone to making mistakes in the evening. I'll come back to it tomorrow with fresher eyes and mind.
I'm not seeing a particularly obvious way to do this. There is already a diode sitting between the 120V rail and the base of Q11. There are no nearby pads on either rail I could exploit for this.
Both R17 and R18 are within tolerances. R17 shows 21.7K and R18 is right at 1.5K.
Similar threads
