2445 power supply failure, cap subs and optoisolator sub question

Hey all. I bought an ebay special that when you turn it on it lights all the LEDs except for the ones on vertical channel buttons. I put a pic of the scope front so you can see what I'm seeing on the front. Checked the voltages and several were a bit low so I guess it's the power supply. Went through troubleshooting chart and it points to the feedback circuit AND unregulated 15 and 5 v sections. I'm hoping to find some help from someone who actually works at Tek in making decisions on subs for the 672D caps and the optoisolator SOC123A. What series caps do I need to replace the 180μF 40v and 250μF 20v  672D Sprague caps with? Also, the optoisolator U1040, pn on the chip is SOC123A. I've seen both MOC1005 and TLP631 TEK L listed on parts lists for that optoisolator part number 156-0885-00 and 156-0885-05.  Can someone help me figure out what to go with? I'm scared if I don't get this right I'll never figure out what is actually wrong with the scope.

Hi Jason,

For the cap replacement question, I would suggest asking the TekScopes mailing list, Tektronix does not offer replacement caps.

For the optoisolator, it's important to know that Tek 9 digit part numbers have a system where the last 2 digits denote backwards compatible (but not necessarily forwards compatible) changes.
So, if your manual lists the optoisolator part number as 156-0885-00, 156-0885-05 should also work.

Has there been an addition to the 156-0885-xx pn? Naturally the TLP631 found in a 2465B is an obsolete part also. As for the tekscopes mailing list there is a fella that sells capacitor kits based on his own reckoning of what they should be replaced with, and a few other posts suggesting Panasonic caps or Nichicon caps of the HE and PW series, and all of the 180 and 250 μF are replaced with 330/50s. I was thinking more along the line of 270/50v and 330/35v Nichicon UHW caps unless there's merit to across the board swap of those two values to 330/50? The fella that sells the cap kits had a Excel sheet showing all of this, but hasn't updated it, and suggests that he's changed some of his recommendations since he posted the sheet which is why I'm iffy about it. Id hit him up, but I have other things to buy from Digikey or Mouser, and not really wanting to pay two different shippings when I could order all I need from one place. Where could I find a RPR that shows the current, or newest specs for the part numbers?:
290-0798-00
290-0800-00
290-0939-00
290-0942-00
156-0885-00
By the way, I have asked a few questions at tekscopes list but I really hoped that maybe I'd get some more educational answers here. For instance, on the optoisolator, the several different parts I'm looking at, the CTR ratings. 20% min on H11A1. 50% min on the TIP117. The third option that I was told is the equivalent of TLP631 has a 100-200% CTR bin. The CTR curve for that device peaks and then has a dropoff as it gets close to it's max If, almost like the CTR is locked in at around 150%. The first two the CTR steadily goes up from 20 and 50% to 600% until it gets to the max If. I'm assuming I don't want to use the binned part and stay with the TIP117 or H11A1. This is the problem, I don't know enough to know whether I'll destabilize something in the feedback by using a different part. But I've asked about that over there and got no response other than the suggestion from an old RPR that one guy had showing the 156-0885-05 (TLP631). Others I've asked have suggested a wrong choice can completely jack up the feedback loop under different loading scenarios. It's definitely something I feel like I'm overthinking, which I'm well known for. I hoped maybe someone who actually knows could help me here. Rein my brain in I suppose.

 

Hi Jason,

There has been an addition but it's also obsolete and no longer available.
Funnily enough, the only 150-0885-xx part still available is 150-0885-00, which you can still buy from us (send an email to buy@tek.com to get a quote).

We have a policy of not giving repair advice besides what's in the service manual, so I'm not going to give you specific recommendations for replacement parts, sorry.

We don't generally publicize specs for our part numbers, so I cannot share those with you, sorry.

Personally, I would say you're massively overthinking it, yes. For the replacement caps, just buy good quality caps that exceed the specs given in the service manuals and you'll be just fine, they really aren't all that critical, especially on those later analog scopes which are surprisingly resilient to power rail noise.

Lol, I figured I was overthinking it. As for the 156-0885-00, I did email parts a while back and when they told me what it cost I thought maybe y'all had reused the pn for something other than a small optoisolator because its pretty eye watering. That's why I've been racking my brain to try and figure this out myself. As an aside, I assume no matter what after power supply repair it would really need to be recalibrated? I'm certain the voltages won't be exactly the same especially if I up the capacitance. Of course with the old caps being -10 +100 tolerance, I'm sure the engineering team accounted for that, so I should be able to easily drop 220-270 μF and 330 μF caps in place of the 180 and 250, with voltages raised as necessary for lead spacing. I saw somewhere that you have to stick to the exact capacitance on two of the 100/25s on the board. Anyhow, thanks for having a look at my comment here and trying to calm my nerves. 73!

Hi Jason,

"Officially", yes, it will need cal after any repair. Unofficially, The 2400s were pretty resilient to power supply fluctuations, you'll probably find it's pretty close to spec.

Why do you want to increase the capacitance on some of the caps? Those supplies were very well designed, if they wanted higher capacitance, they would have put in higher capacitance.

Don't forget that when you increase the capacitance you often also increase other parasitics, those can have unexpected affects. You'll also be changing the resonant points of the filters, which could easily lead to your low pass filter no longer rejecting some of the power supply noise.

I would strongly suggest not changing any of the values unless you have very good reason to do so and understand how it might affect the system.

73

Ah, that's part of the issue I'm having. The suggestions on Tekforums suggests to replace all the 180 and 250 μF -10+100% tolerance caps with 330 μF ±20% tolerance caps. I suppose that was either their judgement based on the original tolerance ratings having a mean point that falls close to 261 to 362.5, or a technical best guess. I really didn't think that would be a great idea for the same reason stated of messing with the resonant points. That's what's causing me such a headache. You can find 180s but it's not easy, 250 is even harder. Seems to be that nowadays 180 and 250 are odd values. It's easier to find 220s and 270s. Then you run into the lead spacing issue, and ripple current ratings. I'm seriously thinking about doing UHW 180s or 220s for the 180, and 270 for the 250s in whatever voltage rating I can find that will fit the lead spacing. Seems like the safest bet to me. I modeled the resonance curve on Elsie changing the two 250s in the pi filter for 5vd to 330 and saw that it moves one of the peaks of the curve before it drops off again to where the peak is very close to 5 khz, where the original values had the 5 khz frequency in the middle of a -40 db "bathtub" (sorry, I can't think of a better term to describe what I'm seeing, you would think someone toting an Extra license would know these things but I'm rusty. Pole zero? Stopband?). I have to think that is very unwanted. Almost like it's asking for problems especially if I'm figuring right on the oscillator frequency for the switcher.  Throwing a peak right on top of oscillator frequency would allow every bit of that 5 khz to randomly pop through depending on loads wouldn't it? 270s move it also, but not as much as the 330s did. That's all I can think of to do, especially with confirmation that messing with the wrong value might be trouble. 180-220 and 270s rated for high ripple and low impedance.

Sorry I misspoke that value. It had to be almost 500μF before the peak of that curve was at 5 khz. At the top of the 100% tolerance end of those 250s. 330μF wasn't as close, they may possibly work. Definitely seems that plenty of room was left by the original engineers for those loose upper tolerances.