Developing a test bench

[richardnhoffman]

Hello everyone,

I have recently started to build a concept test unit for work for testing PC cards. A lot of old test manuals were found that show what we need to have to test each card. Mainly all the boards require power supplies and the rest is test-mesure-adjust-ect. Because I am new to making power supplies and I am on a budget I was thinking of using just linear regulators to make things simple. However, I have come across a few problems.

Most of the boards require a regulated -70 volts. I was thinking of using a lm337 but it does not reach the required voltage. Is it possible and okay to tie a zener or another lm337 to the common of a lm337 with a voltage adjust to bump it to the required voltage? As for references I cannot give any as they are all work associated.

Thank you for you time,
Richard Hoffman

[Eka]

I'd expect that almost any high voltage positive 3 terminal regulator circuit to make a greater voltage output than the regulator can normally handle could be mirrored across ground and adjusted to be used with the lm337. With some, don't short the output. or you will blow the regulator. Others are better protected.

This may give you some ideas on how to go about it: http://www.ti.com/lit/an/snva583/snva583.pdf

[richardnhoffman]

Thank you for the feedback Eka. I need more current, I think around 500ma or even an amp. Does anyone know what I could do to help provide more current but also keep up the regulation? The -70 is going to be constant and is going to be adjusted and set. The rest of the rails I need just a simple lm338.

[Eka]

There are many applications notes from many different manufactures showing how to provide higher current output from positive 3 terminal regulators. Some use op-amps to allow paralleling of multiple regulators, others use FETs or transistors to augment the current handling capability of the 3 terminal regulator. They could provide inspiration on how to modify a negative three terminal regulator circuit. If you have isolated transformer taps for your -70V source it is possible to use a positive 3 terminal regulator to generate a negative voltage by tying the + output to ground, and have the ground be the -70V rail.

Personally I'd just tie the output of two adjustable output 32 or 36 VDC isolated power supplies in series to get the -70VDC rail needed. It likely could be made for under $300 including case. A more expensive solution would be to use a pair of adjustable output 32 or 36 V brick style isolated DC to DC power supply modules in series to get the -70VDC. Other modules could be added to provide the other voltage rails you need. All could be switched at the same time, or in sequence if needed by a simple circuit that controls their enable signals. I power my art pieces with an external 24 to 48VDC power supply. Then I use small 1/8th brick sized modules to make the voltages need to power the electronics. That gets me an art piece with no mains voltages in it.

[David Hess]

The LM317 and LM337 (and LM338 and LM350) are all floating regulators so they *can* be used to regulate high voltages however the input to output voltage differential cannot exceed about 30 volts or they will be damaged which is a problem under short circuit or current limiting conditions.

1/2 amp at -70 volts is not trivial and 1/2 amp through an LM337 limits the input to output voltage depending on the allowed dissipation so a big heat sink will be required.

Figure 11 on page 6 of Linear Technology application note 2 shows the basic idea of using a transistor cascode to turn an LM317 into a high voltage regulator.  Reverse all of the diodes and transistors for the negative version.  A power p-channel MOSFET can be used in place of the PNP darlington transistor.  The minimum output current is increased significantly by biasing of the cascode transistor so take this into account.  If you want to use a zener to set the output voltage, it goes in series with the lower resistor of the output divider.

The only complication is that the LM337 negative regulator is finicky about stability because of its common emitter NPN output stage so pay attention to its output capacitor requirements.

The big improvement I would make to this circuit is to add foldback current limiting.