Author Topic: cap checker repair (Read 5818 times)

pete g · « **on:** January 08, 2019, 05:45:35 pm »

hi all, i'am working a vintage Heathkit IT-11. i'am trying to adjust the current values for the three different types of caps. the electrolytic adjustment pot is adjusted for 2ma, it adjusts fine. however, the mini-electro and the non-electro pots when rotated completely, cannot close the eye tube. the mini-electro needs to be adjusted to 15ua and the non-electro to 2ua. please help.

Nusa · « **Reply #1 on:** January 08, 2019, 06:57:46 pm »

How long since it was known working, if ever?

Any vintage electrolytic or paper caps are probably in need of replacement.

Switch and wiper contacts may need exercising/cleaning.

Tubes may need to be re-seated in sockets. Odds are they still work, but if you have the ability to test them, do so.

Make sure the assembly matches the schematic if you have no personal history with the unit.

pete g · « **Reply #2 on:** January 08, 2019, 10:35:05 pm »

hi. I've owned this piece of equipment since forever. its been awhile since I checked its operation. it was working fine last time I used it, about 6 mos ago.

floobydust · « **Reply #3 on:** January 08, 2019, 11:04:40 pm »

It's 50 years old. Measure a few voltages inside to see if they are good.
IT-11 has very few capacitors. Two series 40uF 350V (to make 20uF 700V) electrolytics which are the first to check. AC ripple on the power supply would upset the Wheatstone bridge which uses AC excitation.

You can measure the voltages on 6BN8 sections V2A through V2C to check if the little capacitors are leaky, or the tube is gassy, while looking at 15uA or 2uA.
The Heathkit IT-11 schematic gives voltages, as well as the manual.

pete g · « **Reply #4 on:** January 08, 2019, 11:22:52 pm »

I checked all the non-electro caps for value and leakage. they check good. the 2 40mfd I could only check at a 100v, they check good for esr and value and leakage at 100v.

mr.fabe · « **Reply #5 on:** January 10, 2019, 03:19:41 am »

I would check the resistors on the tab mount controls AC, AB and also the resistors on the cap type selector switch on the other side of the unit.

pete g · « **Reply #6 on:** January 10, 2019, 03:41:49 am »

I checked those resistors they are in spec.

pete g · « **Reply #7 on:** January 12, 2019, 12:03:11 am »

I replaced both electros and the 6BN8 to no avail. I think its time to fire up the dmm.

pete g · « **Reply #8 on:** January 12, 2019, 06:50:44 pm »

the b plus 180v at the output of volt divider. 6BN8 pin1 minus44v, pin2 minus 12v, pin3 0v, pin6 minus 12v, pin7 97v, pin8 minus6v, pin9 minus 2.4v. 6E5 pin2 126v, pin3 minus43v, pin4 180v, pin5 0v.

floobydust · « **Reply #9 on:** January 12, 2019, 09:08:08 pm »

You have to say what the switches are set to... measurements show the function switch was somewhere like Bridge, so they are not so helpful to troubleshoot leakage function.
Working backwards, the 6E5 magic eye ray-control triode grid voltage is -43V so it is hard cutoff 0º shadow; 0V gives 90º shadow. It should be a few volts negative, around -1V.
6BN8 cathode pin 9 should be +0.45VDC and you are getting -2.4V? That does not make sense unless it is oscillating. Anyhow, the bridge was way off balance when you measured voltages and the tubes are OK.

I would set to a low voltage and measure in leakage test. V2A and the trimpots and both sides of the 1.5MEG R6. Your multimeter will load down the grid side.

I recall you can damage the Heathkit IT-11:
If the cap-under-test is shorted, during a leakage test or
If you short the (+) lead on the capacitor-under-test to ground during leakage test, charging on a HV setting. It cooks some resistors on the voltage switch, it was like R30 down as they are not 10W like R40.

Check R35 47k, discharge resistor R36 680R and 1.5MEG R6.

pete g · « **Reply #10 on:** January 12, 2019, 11:41:50 pm »

r35 good,changed 680ohm/1.5meg checks good.

pete g · « **Reply #11 on:** January 13, 2019, 01:17:06 am »

in leakage position at 25v. on either side of 1.5m--0v. top pot--27v, middle--30v, bottom 50v. 6BN8 pin 7--36v, pin8--0v, pin9--.248v.

floobydust · « **Reply #12 on:** January 13, 2019, 04:37:36 am »

It's hard to troubleshoot a DC-coupled tube amplifier when there is grid current at the 6E5 input.
What voltage on pin 5 6E5 cathode, it connects to... the three trimpots then to GND. So 50V?
Maybe R8 220k high value but it looks like the cal trimpots and resisters are too low in value.
It might just be calibration is set far out. The three trimpots are series, so the top and middle ones (cal) gets thrown out if the bottom one is set way off.

R43 top (paper, mica)
R44 middle (min)
R45 bottom (electrolytic)

I'm saying setting R45 way off to get electrolytic leakage calibrated OK would scuttle the other ranges. It might be your method to CAL the electrolytic range is not quite right.
With power off, you can measure ohms from 6E5 pin 5 to GND for each cap-type and see if the three trimpots are reasonable. I figure trimpot mid-position gives ~22k per type for total ~66k. Just checking the divider string.

pete g · « **Reply #13 on:** January 13, 2019, 02:04:48 pm »

will make checks you request and post back.

pete g · « **Reply #14 on:** January 13, 2019, 05:43:32 pm »

correction: the voltage at the 6bn8 cathode is plus .245v.

floobydust · « **Reply #15 on:** January 13, 2019, 07:57:44 pm »

That means almost all the (6BN8) current is going into the 6E5 grid (36V).
That's why I think the problem is at the 6E5 cathode (50V), where the CAL trimpots are.
Not sure, magic eye tubes are a bit weird to figure out.

pete g · « **Reply #16 on:** January 13, 2019, 09:05:10 pm »

I measured pin 5 6e5 to ground. in electrolytic pos. 0 ohms, min lytic 22.6k same for non-lytic.

floobydust · « **Reply #17 on:** January 13, 2019, 11:34:12 pm »

OK so two of three cal trimpots are at zero ohms which is not right.
I would move the electrolytic from 0->~22k (trimpot mid position) and see if leakage for min lytic works. The top trimpot paper/mica is also zero ohms, put it in the middle.
So 22.6, 45.2, 67.8k is roughly what the resistor string should give for readings. Then I would see what the actual leakage currents need to be, how far off things are.

pete g · « **Reply #18 on:** January 14, 2019, 12:48:28 am »

the top trimpot adjusts the electrolytic setting(2ma). i'am sorry I don't follow. in the electro position the resistance from the 6e5 cathode is 0 ohms. the other two cap types are 22k to ground.

floobydust · « **Reply #19 on:** January 14, 2019, 09:57:29 pm »

I had to draw out the IT-11 Leakage Mode schematic portion to try make sense of it.
I think there is a Heathkit schematic error at the Type switch pin 8 (paper, mica) having nothing connected there. Pin 7 goes to R35 and R36 for the current-sense but I can't find the 680k or so the manual uses for the 2uA range.

The trimpot readings aren't making sense. Does this redrawing help. In the electrolytic position, 6E5 cathode should see R45||R39 around 22k but you are reading zero ohms?

I forgot the IT-11 can dish out 600VDC! In a repair shop I worked at, techs would charge a can-electrolytic to 450VDC and throw it at you yelling "catch!"

I stepped aside but others got a bite on the hand. People are nasty.

pete g · « **Reply #20 on:** January 15, 2019, 12:17:02 am »

I think I made a mistake. I measured again in the electrolytic position and got a reading of 25k from 6e5 k to grnd. sorry. yeah some people have a distorted sense of humor.

pete g · « **Reply #21 on:** January 15, 2019, 12:44:46 am »

are you thinking the pot or pots are bad?

floobydust · « **Reply #22 on:** January 15, 2019, 06:05:11 am »

I thought the 6E5 cathode voltage is too high, so the trimpot string might have a bad resistor. An easier way might be to spend a minute and check their resistance. You get 22k across the R45//R39 combo which is good, just need to check the other two.

The current-sense resistors could be off value. You replaced R35 or R36? R35 47k will smoke if a shorted cap gets leakage tested at HV, and R36 680R is also used to discharge caps so it gets hit hard. Both parts should probably be a few watts to be tough.

I would confirm R35 and not sure who the last sense-resistor (paper, mica) is on pin 8.

Surprised to learn the 6E5 magic eye tube and patent go back to mid-1930's.

pete g · « **Reply #23 on:** January 15, 2019, 07:44:48 pm »

componets changed: the two main electros, r6, r9, r36, r37, r38,r39, c3, c11, 6bn8. problem remains. ugh!

floobydust · « **Reply #24 on:** January 16, 2019, 04:22:38 am »

See if the IT-11 works reading a resistor value. Put a couple same value resistors (ref. and test) and see if the (bridge) works.

Sometimes I've seen poor soldering, cold solder joints in Heathkit builds. They were kits after all.

If resistance measurements work, I would go on and troubleshoot one leakage range like miniature electrolytic. Assuming electrolytic leakage is working OK with 2mA. Follow the cal procedure with 15uA and measure the tube voltages. 6E5 cathode pin 5, grid pin 3 and 6BN8 plate pin 7, grid pin 8, cathode pin 9.

The only parts left are C4 0.5uF, C5 0.1uF, C6 2uF and the range switch parts. Not much more to this beast.

pete g · « **Reply #25 on:** January 16, 2019, 03:40:26 pm »

hi, yes, there's not much left. wish I could find a bad part. because of all the parts I changed, not one tested bad.

Nusa · « **Reply #26 on:** January 16, 2019, 05:17:40 pm »

You could try measuring contact resistance on various switches, if you haven't. If you get any measurable resistance on closed contacts, you have probably found a problem.

Armadillo · « **Reply #27 on:** January 16, 2019, 09:09:28 pm »

Won't it be easier to know the actual voltage/current readings?

1) Are you using the with milliampere/uampere technique or without?

2) Under Voltage Test - Please provide all the 16 voltages as measured with the voltage meter at the Test Terminals.

with that at least we have a better foundation to start tracing.

pete g · « **Reply #28 on:** January 16, 2019, 11:45:58 pm »

I was scoping the line that feeds the grid of the bn8, ans saw a lot of noise with big spikes intermittently. trying to localize, I disconnected pin 6 and 10 of the B-D-L switch. need to find the source any ideas?

pete g · « **Reply #29 on:** January 17, 2019, 12:08:03 am »

will do that and post back.

pete g · « **Reply #30 on:** January 17, 2019, 12:10:52 am »

not using a milliameter. will check voltages.

floobydust · « **Reply #31 on:** January 17, 2019, 06:28:59 pm »

6BN8 grid is a high-impedance node in leakage test, it should be sensitive but quiet.
There is nothing really for noise filtering for the 6BN8 aside from R7 100R so I'm not sure how it does with modern RFI. Move your cellphone or Wi-Fi stuff.

I would suspect an open switch contact. You can measure ohms from 6BN8 grid through the rotary switches.
I use a small pencil eraser to clean oxide off the wiper and gently drag paper through the wafer contacts to clean them, or use contact cleaner spray.

pete g · « **Reply #32 on:** January 17, 2019, 07:31:23 pm »

I hung a 20ufd cap on the line that feds the grid of the bn8 and that seems to have solved the problem as far as setting the calibration of the min lytic position. the mica,etc. cal is still not working. sometimes I can get the eye to close and then it opens again?

floobydust · « **Reply #33 on:** January 19, 2019, 06:25:33 am »

Putting a capacitor at the 6BN8 grid will filter hum and noise there for leakage-test mode.
But that part's leakage current alone will upset tests and it will kill the bridge mode for measuring capacitance because that uses AC. So I don't think it's a fix.
I would not scope the 6BN8 grid, but instead the plate if looking for hum or oscillating.
If it's a problem, the 6E5 you can pull and see if it's C4 or C11 are making it worse.

You are sure the new filter caps C1, C2 are working, at least 20uF 700V?

I would check the grounding, from the line cord to chassis and R41/R42 and the ground on the 200V side. It also goes to the bridge transformer. It's a few minutes for continuity tests.

The paper, mica leakage test, I still can't find the current-sense resistor. Thought it would be at pin 8 on the rotary switch.

Beyond this, need some pictures inside.

pete g · « **Reply #34 on:** January 19, 2019, 02:49:01 pm »

there is noise on the plate of bn8 sometimes reaching 1-2vpp. I did find a wiring error to the cal trimpots and corrected it. I would think the solution to this problem is being told by the voltage readings on the bn8. I think if they can be analyzed, the problem would be solved. what do you think? i'am able to calibrate the minllytic and mica,etc, but the adjustment won't hold the eye tube opens again or flutters.

pete g · « **Reply #35 on:** January 21, 2019, 01:45:53 am »

hi guys, i'am going to shelve this project for now. wanted to take a minute, to thank those folks who contributed. much appreciated.

pete g · « **Reply #36 on:** January 26, 2019, 03:11:14 am »

hi, the problem with this tester, turned out to be off the wall. after years of me dousing the phenolic switches with contact cleaner, they have become conductive and leak voltage erratically. never knew this could happen, I guess the moral of this drama is not to douse those switches.

Nusa · « **Reply #37 on:** January 26, 2019, 06:15:12 am »

Glad you finally spotted something! How are you going to fix it? Slide switches are easy enough, but those wafer switches are likely unobtainium.

pete g · « **Reply #38 on:** January 26, 2019, 02:11:01 pm »

I think its curtains for this meter. but i'am going to use some of the parts to build another one.

TERRA Operative · « **Reply #39 on:** January 26, 2019, 02:54:45 pm »

Can you pull the leaky switches out, clean them thoroughly in isopropyl alcohol and bake them in an oven to completely dry them out?

Failing that, maybe modern replacements could be shoe-horned in?

Armadillo · « **Reply #40 on:** January 26, 2019, 03:14:42 pm »

The 16 readings would have exposed the leakage switches, yes?

Quote from: Armadillo on January 16, 2019, 09:09:28 pm

2) Under Voltage Test - Please provide all the 16 voltages as measured with the voltage meter at the Test Terminals.

pete g · « **Reply #41 on:** January 26, 2019, 10:06:05 pm »

baking it might be worth a try. but, I wonder if it would get really brittle. I don't think alcohol is s good choice as a solvent.

pete g · « **Reply #42 on:** January 26, 2019, 10:07:47 pm »

the voltage readings were not affected, but the wafer did start to smoke at 600v.


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: cap checker repair (Read 5818 times)

Share me