Oscilloscope (DSO) and bench PSU choice for repair/analysis...

For discussion of triage and repairs to SGIs as well as general hardware questions
User avatar
jan-jaap
Posts: 80
Joined: Fri Jun 01, 2018 5:11 pm
Location: Netherlands
Contact:

Re: Oscilloscope (DSO) and bench PSU choice for repair/analysis...

Post by jan-jaap » Thu Aug 02, 2018 4:53 am

I don't think you need to go overboard with the bandwidth of the scope if all you do is testing PSUs. I've got an old 40MHz Tektronix and it works just fine, the (chopper) frequencies inside a PSU are in the dozens of kHz to maybe a couple of hundred kHz.

What's your intention with a bench PSU? If you want to run an SGI system off it, you need lots of amps @ low voltages, e.g. a lowly Indigo2 PSU puts out 7 different voltages, including 40A @ 5V + 48A @ 3V3. That's not practical with a bench PSU, especially if you probably have at least one working actual SGI PSU. (At work) we more commonly use bench PSUs to bring up new hardware, where you want to limit current in order to minimize the damage when there are bugs in the design...

User avatar
mapesdhs
Posts: 10
Joined: Tue May 29, 2018 8:22 am

Re: Oscilloscope (DSO) and bench PSU choice for repair/analysis...

Post by mapesdhs » Thu Aug 02, 2018 8:59 am

jan-jaap wrote:
Thu Aug 02, 2018 4:53 am
I don't think you need to go overboard with the bandwidth of the scope if all you do is testing PSUs. ...
Heavens no, I only mentioned PSUs because that's the easiest type of faulty item I can deal with to get used to all the equipment, learn the methods, etc. (and I have lots of them so if I initially screwup then it's not a big deal, whereas beginning all this by trying to fix a DM6 would be most unwise), along with items like TRAM modules (I have more than two dozen bad TRAMs). I intend fixing everything from PSUs to all types of board for 4D, PI, Crimson, Challenge, Onyx, Indigo, Octane, O2, the lot, and the option cards & suchlike. Everything. Or at least, anything that can be fixed within the limits of however much I can learn. I want to be able to do things like check what's going on on a VPro board, a DM6, DCD, etc. It does though need to follow demand, and fixing PSUs is very relevant indeed. I started all this because I had to tell a hospital in Mexico because I was not able to repair the PSU from their Personal IRIS.

And in time I may take on repairing PC stuff too, who knows. Actually already did one, a basic SMC repair of an ASUS Rampage IV Extreme.

jan-jaap wrote:
Thu Aug 02, 2018 4:53 am
What's your intention with a bench PSU? If you want to run an SGI system off it, ...
No, that's not the idea. I assumed you knew the intent is to be able to check circuit functionality, detect shorts, bad ICs, etc. One could certainly use a bench PSU to act as a sustitute if some part of a normal PSU is bad, but that would be limited to probably just Indy and O2, maybe Octane at a stretch depending on the model. One realises very quickly that the typical output power rating of a bench PSU is pretty low (higher power units are expensive), and of course the efficiency of linear types can be very low indeed. I suppose one could use multiple units to shoehorn something together, but as you say, some systems need very high currents of certain low voltages, and if you think I2 is bad then check the kind of currents flying around an Onyx. :D I'm sure you already know.

Hmm, I suppose this is maybe why articles about converting ATX PSUs into bench PSUs are popular, such units could provide much higher currents. Not something I would do though, but I guess one could use an ATX PSU as a substitute for a particular bad feed in an SGI PSU (as long as the voltages match), but as you say, often there are unusual voltages. Either way, this isn't really as important as the initial stage of just getting to grips with how SGI PSUs work, the circuit layout, what tends to go wrong with each, establishing the diode-mode test response behaviour of a working unit as a baseline, etc. There's so much to do; it's like standing at the bottom of K2. :D

jan-jaap wrote:
Thu Aug 02, 2018 4:53 am
where you want to limit current in order to minimize the damage when there are bugs in the design...
Exactly, it's about testing and fault finding. I'm just at the very beginning of this.

BK 393 DMM arrived just now. 8)

Ian.

Image

User avatar
mapesdhs
Posts: 10
Joined: Tue May 29, 2018 8:22 am

Re: Oscilloscope (DSO) and bench PSU choice for repair/analysis...

Post by mapesdhs » Thu Aug 02, 2018 9:22 am

PS. I've ordered the 1204X-E. :) Not yet decided on the PSU.

Post Reply