DEC power supply 725 going hot in TU56H

[anders_bzn]

I'm about to restore a TU56H Dectape drive. So far I have been cleaned it up a bit, replaced the motor capacitors. I have slowly reformed the big capacitors in the power supply and yesterday I tried it for the first time. I got really hot!

Looking at the schematics there are a 100 Ohms resistor in parallel with the big bulk capacitors on both power rails. The one on the 38V rail got really hot! Is this normal?

Schematic: http://www.bitsavers.org/pdf/dec/dec...nual_Jan73.pdf on page 71 (pdf-number).

I measured the voltage on the rail to 43 Volts, this gives a power dissipation of 18W! I'm not surprised that it gets hot! I guess that the voltage will drop a bit when the supply is loaded, but with the nominal 38 Volts it will generate almost 15W of heat...

[m_thompson]

What will you connect the TU56 to?

[anders_bzn]

I'll connect it to my small 8/a: http://www.pdp-9.net/pdp-8a-100 using the TD8E interface for a start. The thing is that the drive and interface is a loan from Pontus, but I'll use it to debug the G888 replacement I have been working on for years.



The card is basically working, Mattis Lind has tested it and booted some OS (don't remember which) on a PDP-11. But it didn't pass the maindecs yet.

[anders_bzn]

The really long term goal is to build my own tape controller so I can dump the tapes that I got along with my PDP-9

[m_thompson]

I wrote some simple diagnostics to help debug a TD8-E.

[Al Kossow]

The really long term goal is to build my own tape controller so I can dump the tapes that I got along with my PDP-9

A simple interface isn't difficult to make to read tapes. I built one about 10 years ago that just connected to a PCI parallel port card. You clock the mark and data tracks with the clock track.
I just uploaded the dec and linc tape decoder code to http://bitsavers.org/projects/dectape
This was written to interface to a PCI parallel port card in a Mac G3 running OS9, so there is a lot of code just for dealing with finding the card on the PCI bus.
The main thing of interest are the guts of decoding the DECtape frames.

[m_thompson]

Al,

Did you use the same clock signal for the DECtapes and LINCtapes?

[Al Kossow]

Al,

Did you use the same clock signal for the DECtapes and LINCtapes?

Yes, you use the clock track off of the tape. the whole system is self-clocking.

You have to do this, because the phase-encoded data cells aren't at a constant clock rate because the tape doesn't move at
a constant rate.

Also, dectape preamps are designed to self-oscillate until they are up to speed, so you need to ignore the inputs for a while.
You can see this if you watch the clock track on a scope. The up-to-speed circuit inside the dectape controllers count on this
behavior to work.

[Al Kossow]

The one on the 38V rail got really hot! Is this normal?

No. One of your motor driver transistors is probably shorted.

[m_thompson]

Al,

I asked because the PDP-12 tape controller inverts the clock signal when it is reading DECtapes.
The TU56 has both polarities of clock signal available.
Did you spin the LINCtapes in their normal direction, opposite of a DECtape?