Author Topic: No luck on my first attempt at acquiring a cesium frequency standard (Read 55152 times)

Vgkid · « **Reply #75 on:** March 17, 2016, 09:11:11 pm »

About those yellow axial caps

Another option for those snap caps, would be switching to radial. With the appropriate lead spacing, useful for those snaps with multiple mounting leads. plus a lot more common.

hydrogen maser · « **Reply #76 on:** March 17, 2016, 09:17:58 pm »

The yellow axials (6 of them) are all 20uh inductors. edpalmer42 - good tip on getting higher voltage caps to fix the size issue, I was wondering what I was going to do about that. It sure could not hurt to open up the power supply and take a peek, all you have to do is take out the screws on the top plate then about 6 screws in the board and tilt it up so the left side is down. There is a flex cable that ties that board to a lower board so just keep that in mind. Vgkid, you did not say the yellow ones were caps, you just said they should be glued down and they are

Good tip on the radial, I can put anything in there as long as it is the same footprint or smaller, but I like edpalmer42's idea of going with a higher voltage rating, can't hurt.

edpalmer42 · « **Reply #77 on:** March 17, 2016, 09:26:21 pm »

Look very carefully at those capacitors. They might not be bad after all. When I first saw the picture, I didn't think "leaking capacitors", I thought "water (or humidity) damage". What do you see if you look at the input power lead on the ion pump supply with your scope? How much ripple? Maybe there is another fault in the power supply.

Ed

hydrogen maser · « **Reply #78 on:** March 17, 2016, 09:39:17 pm »

I am going to replace the caps before I do anything else. At least 6 of those caps have pretty serious corrosion going on - white powder on the top of them, discolored and lumpy, you can feel the bumpy surface on them. There may well be other problems with the power supply or other parts of the system but I am going to replace the caps first, they are not very expensive and then see what it looks like.

Vgkid · « **Reply #79 on:** March 17, 2016, 10:00:43 pm »

I will usually overspec my caps when i replace them. Usually keeping the capacity the same, but sometimes 1.5-2X larger. Mainly if there is alot of ripple still present.

edpalmer42 · « **Reply #80 on:** March 17, 2016, 10:16:35 pm »

I took a look at my power supply board. I remember now that I decided not to pull it because to do so, you have to remove the 5045A from the chassis to get at 4 screws on the side that attach the heatsinks to the chassis. At that time, I had no reason to remove it. Luckily, my board looks fine. Date codes on the capacitors are 93 or 94. One odd thing is that even though there is a bit of glue to supposedly hold down the capacitors, every one of them is spaced away from the board so they can wiggle and vibrate. Of course, the glue is broken now. Not impressed.

Ed

hydrogen maser · « **Reply #81 on:** March 17, 2016, 10:35:12 pm »

edpalmer42, your main board is different from mine. I see the STEL-1173 NCO on your board is in a socket. Did you install the socket or did it come that way?

edpalmer42 · « **Reply #82 on:** March 17, 2016, 10:41:39 pm »

I was just going to post this message, but your question got in first. I guess we're thinking in lock step.

By the way, if you look in the background of the picture above, you can see half of an IC in a socket. That's the STEL-1173 that typically fails with age. I installed the socket so that if it fails again, replacement will be trivial. If you look really carefully you can see a row of three surface-mount capacitors near the socket. There's just room for a piece of paper between the closest capacitor and the socket!

Also, note the little 'mezzanine' board behind the socket. HM, I didn't see this in your pictures. It looks like they did a circuit mod and rather than hack and slash, they just abandoned the board and added a new one above where the old components were. They used really stiff wires to connect up the new board to the main board and didn't need any more support.

Ed

hydrogen maser · « **Reply #83 on:** March 18, 2016, 03:06:49 pm »

The glue is really old and brittle on the caps, just wiggling them a little back and forth after desoldering breaks the glue bond and they are coming right out. Took a look at everything else and don't see any other obvious problems. bad power supply explains why basically nothing was working other than the oscillator, there must have been enough voltage and current for that to work. After I get new caps in it will be interesting to see what happens.

hydrogen maser · « **Reply #84 on:** March 18, 2016, 04:01:28 pm »

Yeah, I think these qualify as "bad caps". Some of them were so far gone that the pins just fell out of them during removal.

edpalmer42 · « **Reply #85 on:** March 18, 2016, 04:35:14 pm »

What a mess. I wonder why yours ended up in such bad shape compared to mine. My caps are the same brand and series as yours, but they're about four years older.

Ed

hydrogen maser · « **Reply #86 on:** March 18, 2016, 05:05:56 pm »

Danno, perhaps it was in a harsh environment, high ambient temp, not enough circulation? I don't know the history of this device so who knows what has happened to it. Perhaps the PS went bad and then it was stored for years without any climate control? Not a single one of those caps is any good. Well, ONE of them looks ok, the one on the far left in the lower row but I'm not going to even bother testing it, no way I would trust it. Edit: OK so I did test it just to amuse myself - the one cap that looks ok reads just fine on the tester - capacity good, ESR less than 1/10 of an ohm, delta good. Sampled a few of the rest of them and they read what you would expect: little to no capacity and very high esr - hundreds of ohms and some in the kilo ohms. I wonder if that one little cap was doing all the work in the PS. The little cap that could. Shame it has to go to the graveyard with it's friends after putting up a valiant effort. Actually going to keep that little guy around and put him in a place of honor.

Macbeth · « **Reply #87 on:** March 18, 2016, 06:42:30 pm »

Quote from: edpalmer42 on March 18, 2016, 04:35:14 pm

What a mess. I wonder why yours ended up in such bad shape compared to mine. My caps are the same brand and series as yours, but they're about four years older.

Ed

The date code of late '98 is awfully close to the cursed Capacitor plague that appeared in '99.

hydrogen maser · « **Reply #88 on:** March 18, 2016, 08:04:36 pm »

Hm, this doesn't seem right - there is continuity across some of the capacitor holes on this board. I would not expect that to be the case. Unless some power supplies do that for some reason?

Macbeth · « **Reply #89 on:** March 18, 2016, 08:11:23 pm »

Quote from: hydrogen maser on March 18, 2016, 08:04:36 pm

Hm, this doesn't seem right - there is continuity across some of the capacitor holes on this board. I would not expect that to be the case. Unless some power supplies do that for some reason?

Continuity like very low ohms? Perhaps the rectifier diodes have shorted and you are measuring the transformer secondaries? Of course with no PSU circuit diagram to go on it's just a hunch. Easy enough to check.

hydrogen maser · « **Reply #90 on:** March 18, 2016, 08:27:27 pm »

Good call! D20 is shorted, pulling it out now.

hydrogen maser · « **Reply #91 on:** March 18, 2016, 08:42:01 pm »

diode is fine, something else in the circuit in parallel with the diode is shorted.

Macbeth · « **Reply #92 on:** March 18, 2016, 08:48:01 pm »

Quote from: hydrogen maser on March 18, 2016, 08:42:01 pm

diode is fine, something else in the circuit in parallel with the diode is shorted.

...or in series, if D20 is actually a Zener

hydrogen maser · « **Reply #93 on:** March 18, 2016, 08:55:55 pm »

it is a multi layer board with the top and bottom layers being ground planes. Oh well, it was fun while it lasted.

hydrogen maser · « **Reply #94 on:** March 18, 2016, 09:22:54 pm »

This thing is weird. 4 caps shorted, but not all to each other. 2 caps (4 pads) have continuity between them as do another 2 sets of caps but there is no continuity between the two sets of shorted caps. I am tempted to just put new caps in there and if it blows up then it blows up. What do I have to lose at this point?

Vgkid · « **Reply #95 on:** March 18, 2016, 09:24:57 pm »

Could you try powering up the circuit externally?

hydrogen maser · « **Reply #96 on:** March 18, 2016, 09:47:22 pm »

It is all integrated together, there are basically 3 boards (not including stuff on the bottom). The mostly power board has a 20(ish) pin flex connector to a lower board (soldered on both ends) with a bunch of multipin connectors coming off that go everywhere else. There is even an RF connector on the "power" board. I guess I could disconnect everything from the bottom board and the RF from the "power" board, put new caps in it and let er rip. If it blows up then at least that top board was toast anyway and it won't hurt anything else other than itself and the lower board. That would spare the main board and all the other stuff from any harm. But if I do that and power it up and the magic smoke does not escape, that does not mean it will be ok to plug everything back in. Hm, what to do....

Vgkid · « **Reply #97 on:** March 18, 2016, 10:33:10 pm »

That doesn't make it easy. I was hoping that you could have been able to remove the power board, and power everything up from its connector.

hydrogen maser · « **Reply #98 on:** March 18, 2016, 10:35:33 pm »

Going to sleep on it and come up with a plan for how to proceed tomorrow. For all I know everything else works fine and I don't want to damage anything else. In the worst case scenario where I can't fix or determine why those caps are shorted I will proceed with replacing the caps in the "power" board, isolate the two small boards from everything else and power them from my lab power supply with current limiting and slowly turn up the current and see what happens. If they aren't fixable then I can wait for another broken 5045A to appear and hope I can scavenge parts and piece together something. From what I can tell the 4040A PS and electronics are ok (

?). No need to rush this and break something beyond repair. I welcome any suggestions anyone has on how to proceed.

edpalmer42 · « **Reply #99 on:** March 19, 2016, 04:39:01 am »

I don't know if this will help, but I measured the voltage on both leads of the 15 capacitors. There were a few odd values as you'll see. I looked at each lead with the scope and there was no significant ripple or signal on any of them. A few of the capacitors had component designations on the board, but most didn't so I added my own and included a picture with my labels.

C3 C4 C18 CA C20 C19 CB CC CD CE CF CG
Minus Lead +77.5mv +77.5mv +17.60 0 0 0 0 0 -15.19 -15.22 0 +17.83
Plus Lead +32.3 +32.3 +28.74 +5.22 +15.12 +25.80 +5.16 +25.76 0 0 +15.10 +18.72

CH CI CJ
Minus Lead 0 0 +17.63
Plus Lead +4.96 +4.99 +18.76

All voltages were measured with respect to the chassis. The minus leads of C3 and C4 are definitely not zero. The others were within a few millivolts of zero so that's what I showed. I double-checked the voltages on CG and CJ. They're correct. Very suspicious.

Based on what my unit reports for voltages, I'm making the following assumptions:

+ 5 volts: CA, CB, CH, CI
+15 volts: C20, CF
-15 volts: CD, CE
+24 volts: C19, CC
System Voltage: C3, C4

No clue what C18, CG and CJ are for.

Since mine is running right now (I'm doing a data run to compare the Cesium to my Rubidium house standard), I can't do resistance measurements to confirm that these capacitors are actually paralleled as I've shown. I wouldn't be surprised to find multiple, independent outputs with the same voltage. Check yours and correct as necessary.

Ed


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

Author Topic: No luck on my first attempt at acquiring a cesium frequency standard (Read 55152 times)

Share me