Ideal bench power supply

[calexanian]

After the blog with free electron where the virtues of power supply shunting were discussed I started thinking about what kind of supply I might want for my new bench I am building at home. Not many supplies under $300 do that. Some have clamp circuits but none that I could find have something that will really regulate. Many of the circuits I am working with have heavy solenoids in them and they can send noise and junk back into the power supply and the idea of a source/shunt combo supply is an attractive one. In addition to this I was looking at a tube of LM3886's I have and thinking about their internal overcurrent and thermal protection. An idea is brewing......

Thoughts?

[pinkysbrein]

How would you parallel them?

[mian2zi3]

I didn't see the free electron blog.  Anyone have a reference?

[calexanian]

I am sorry. Reference cross up. The mention was in the recent AmpHour. Dave and Vincent had a brief discussion about regulators that shunt reflected voltage spikes from a circuit And no need to parallel. A single device would be more than adequate for a simple supply. I have used 3886's as regulators before. The two limiting factors are the output short current and the thermal shutdown which is more a factor of heat dissipation and proper heatsink design and unregulated rail voltage scaling (Switching taps, or power chopper)

Sorry I was not more specific.

[TMM]

How often do you actually need your PSU to sink current and how much current overhead does your PSU have? On the odd occasion i need mine to sink current i'll just put a power resistor across the psu output. Then you can sink as much current as the resistor is drawing.

[calexanian]

We drive solenoid and motor loads, not to mention voltage boosters that create a lot of noise up to about 4 amps. Most of our benches have 3 amp supplies but we have a couple larger ones floating around the shop. Remember, this is a theoretical exercise of an ideal supply.

[uoficowboy]

How often do you actually need your PSU to sink current and how much current overhead does your PSU have? On the odd occasion i need mine to sink current i'll just put a power resistor across the psu output. Then you can sink as much current as the resistor is drawing.

Your solution can be effective, but it's of course far from ideal.

There are four quadrant supplies out there - the Kepco BOP line is quite nice. They use noisy AC fans that are always on, however. Built like tanks. Not ideal for a lab setting. Lambda makes the BOSS line, but I believe they suffer from the same noise problem. Keithley makes sourcemeters which can sink and source current and are fairly quiet. Also very expensive - $2K+.

[robrenz]

There are four quadrant supplies out there - the Kepco BOP line is quite nice. They use noisy AC fans that are always on, however. Built like tanks. Not ideal for a lab setting. Lambda makes the BOSS line, but I believe they suffer from the same noise problem. Keithley makes sourcemeters which can sink and source current and are fairly quiet. Also very expensive - $2K+.

The BOSS line is no longer made but I have two NOS 20V 10A units. They are extremely noisy but very versatile. They can act as a high power op-amp up to about 20kHz. Makes a nice amplifier for a FG.  My thread on mine here.  My Keithley 2450 SMU is silent but only sources and sinks 1A.

[uoficowboy]

There are four quadrant supplies out there - the Kepco BOP line is quite nice. They use noisy AC fans that are always on, however. Built like tanks. Not ideal for a lab setting. Lambda makes the BOSS line, but I believe they suffer from the same noise problem. Keithley makes sourcemeters which can sink and source current and are fairly quiet. Also very expensive - $2K+.

The BOSS line is no longer made but I have two NOS 20V 10A units. They are extremely noisy but very versatile. They can act as a high power op-amp up to about 20kHz. Makes a nice amplifier for a FG.  My thread on mine here.  My Keithley 2450 SMU is silent but only sources and sinks 1A.

I have used a Keithley source meter (2602B). Awesome specifications, but horrendous interface. Ugh. I think it really is designed to be used from a computer or with scripts, instead of as a normal bench instrument.

I personally own a HP 6825A. That's +-20V at +-2A. It does not have a fan, so it is dead silent. It can act as a four quadrant voltage source, or as an amplifier. It can do fixed gain or variable gain amplification. Very handy little device. I picked it up for about $100 on Ebay but it was very broken at the time. After a decent number of repairs it is now fully operational. It has an absolute current limit (ie you can set the magnitude of the current limit). This is both good and bad. For most people it is good. but I did like how on the Kepco BOP device I used to have you could have it act as a current amplifier rather than a voltage amplifier (so you'd set a current, not a current limit). So the 6825A is definitely missing some functionality IMHO. Even with that limitation, the 6825A is among my favorite pieces of equipment that I own.

They also made the 6826A (+-50V, +-1A) and 6827A (+-100V, +-0.5A). All three devices are pretty rare on Ebay, but you'll see a good deal occasionally...