So i recently scored a Agilent 6032A system power supply and had to have a look inside at what makes it tick.
The main reason for poking inside is to see if something could be done by the incredibly loud fan and if there is any possibility of front panel connectors. I quickly found the answer to that was no and no. The fan is actually a AC fan that runs from mains so that would explain the massive amount of airflow trough the unit. And while most of the front panel is blank it turns out that the PCB behind it actualy fills the entire area in order to fit in all the DIP chips with the logic for multiplexing the front panel.
But none the less the insides are pretty interesting as pretty much everything inside is massive and overbuilt. Just have a look at the size of that current shunt! Also as one would expect the caps in here are top notch with all the big ones being Cornel Dublier and the small ones being Nichicon. Even the front panel switch doesn't directly switch power to it. It instead switches two big ass relays that then switch live and neutral to the unit.
Another surprising thing is that i can not find a single spec of dust in here. The unit must have run inside a clean room or has sat on a shelf. I gave the pretty toasty input capacitor bleeder resistors a sniff after running it for a while and i could notice a smell off them so its possible this is a almost new unit. The thing also really didn't want to come apart as they used thread locker by the truck load on the internal screws. I didn't take it completely to bits because its not all that easy to take apart and i can see pretty much all the interesting bits.
I wonder how this compares to modern high power system power supplies that are made today.
I have the older 6012A model (actually 2 of them) and put a fan temperature contol board in them with an NTC sensor on the heatsink. I have the fan running slow in mine and throttling up from 30 degrees to full power when the heatsink reaches 50 degrees. I have not managed to make the fan run faster during use!
I also have an HP 6572A which follows the same recipy as the 6012A power supplies but it uses more modern components. For later models it seems Keysight rebrands TDK-Lambda PSUs.
The heatsinks do indeed stay cold, i am more worried about the power resistors scattered around the place. Running it without the cover on did make the bleeder resistors get quite hot. My plan so far is to simply replace the fan with a regular 12V computer fan that could be easily speed controlled, but i know it will never move anywhere near as much air as this AC monster of a fan does. But then again i don't see why in the world would i run this supply at full load in the first place.
I too have one of the older 6012A power supplies and it is great. I have a 6011A (0-20v, 0-120A upto 1kW) that is in the "to be repaired"pile. Big stud mount rectifier diodes on the output need to be replaced. There doesn't seem to be much difference between the 6011A and 6031A power supplies, but I am sure there are deep down. I also like the seemingly modular design. The main PCB looks similar across the models, then they choose the transformer they need, the plug in modules and a few other bits and pieces and they are done.
Are you able to take some photos of the 2 PCB's on those heatsinks? They should just unplug and lift out.
Yeah a lot of the modules are reused. The so called FET module is the same across almost all of them. But the transformers and diode boards change. So i would assume its very easy to make one working unit out of two broken ones as long as they died a slightly different fate.
I been looking trough the service manual for ideas of how to retrofit a quieter fan. Luckily it contains full schematics and there is a lot of them as this is quite a complicated unit. Turns out that the fan is actually running from 110V AC that is pulled from one of the dual primary windings (220/110) of the mains transformer that runs the control circuitry, That way it always gets 110V no matter the mains voltage. Quite clever but annoying when you want to put in a 12V fan. Luckily the fan is a standard 120mm fan so a new one can simply be bolted on.
What i could do is take the guts from a 12V switchmode wallwart and use it to turn the 110V AC in to 12V, but i was looking for a more elegant solution. Looking trough the schematics the i think the best rail to take it from is the one called "+5V UNREG" This comes directly rectified from the transformer and sits at 21V. Its so high because a switchmode converter is used to turn it in to 5V to run some digital logic. It sounds to me like this rail has a fair bit of current behind it and since it runs a switchmode 5V regulator it means that noise is not an issue on it.
I will take the control boards out since that's where i can tap off the rail i need. Plan is to make a tiny board that will turn the 21V in to about 10-18V depending on how hot the heatsink is, feeding that in to a standard 12V PC fan.
Also here is the service manual if anyone is interested(Schematics start at page 113):
http://cp.literature.agilent.com/litweb/pdf/5959-3344.pdf
You can also use a simple triac based dimmer and put it in series with the existing fan. That is how my fan control board works.
Not all AC motors like speed reduction by a simple dimmer. Some might even overheat.
You can also use a simple triac based dimmer and put it in series with the existing fan. That is how my fan control board works.
How well does that work? It was my initial idea, but i decided against it since looks like an asynchronous ac motor. They usually don't work all that well with dimmers and i would assume the fan is still loud even at a low rpm due to how these motors shake around at the line frequency.(It has a pretty expressed low tone noise after it gets up to speed, that doesn't sound like blade noise but rather motor noise)
Not all AC motors like speed reduction by a simple dimmer. Some might even overheat.
It works well with AC fans typically found in equipment. They (usually) are impedance protected to make sure they don't overheat with the impeller blocked.
As promised here are the CPU and control boards.
Here is also the FET board and the fan.
First thing i noticed about the fan is that it is really heavy, later on i noticed why. The fan and the fan case is actually die cast metal! No wonder the fan spins forever when you turn it off.
I was also investigating how to make the fan quieter. I didn't have a triac dimmer on hand so i just used some power resistors to drop the voltage. Turns out that 500 Ohm is about right for cutting the speed in half.
But the problem was that even tho the fan was running slower it still seamed to make just as much noise, after some fooling around i found out that whats actually causing the noise is the FET board heatsink being too close to the fan, causing a noise as the blades went by. The humming tone would clearly follow the fans RPM. Since i was out of ideas i tried flipping the fan around so that it would blow out the back like most bench instruments. This made a BIG diference to the noise, suddenly there was no more humming noise, the air wooshing noise was the dominant sound as it should be. As a bonus it also got rid of the weird feeling of air blowing out around all the buttons on the front panel.
At the moment i was too lazy to add temperature control to it and the fan was still moving quite a lot of air so i decided to permanently install the mod. I added some connectors to the resistors and heatshrunk everything. Then to mount it i drilled a 3mm hole in the side of the internal case to bolt the resistors to the case to help cool them (They get pretty toasty as they have to dissipate roughly 5W).
I tried to make the mod look like something that was originally there since i love how well layed out and tidy the internal construction is from the factory.
Good idea flipping the fan around since the noise was from turbulence between the fan blades and the heatsink fins. I'll have to keep that in mind for future fan noise situations. From the photos, it looks like air should still flow through the heatsink in the reversed setup. If not, you could always add some plastic shrouding to help direct the airflow.
Usually, fans are set up to blow into the enclosure to increase the internal air pressure thereby preventing dust from getting sucked in from every little opening in the panel, cover, etc. Thus, an air filter just needs to be installed at the fan inlet. Just something to be aware of, depending on the environment you're going to have your supply running in.
I've been casually looking at high-current supplies for a somewhat esoteric use. In order to fully adjust my electronic load (60V, 60A, 300W), I need a variety of voltage/current combinations from 62V@0.5A and 50V@6.2A to 17V@20A and 5V@75A. This unit almost covers it and at a reasonable cost.
Linear HP supplies that I've used have quite a bit of headroom above the quoted specs. Do you know if that's the case with this SMPS model? If so, how much (e.g., how far above 60V or 50A will it go)?
I think they go a few volts/amps over the maximum rating. My HP6012A didn't complain when loaded with 1350W for a few seconds. I got two of them so I can connect them in series for higher voltages and parallel for more current.
Yeah, I may end up going the two-supply or dual-channel route. It's more flexible, anyway.
I think it can go to 52A before it says error. But you can just as easily parallel any PSU in constant current mode to give you the few amps short of the 75A you want. But if you buy two of these system PSUs they can be made to follow each other by wiring together the right pins on the back.
Sucking air from the back does make sense from a dust point of view, but these power supplies don't come fitted with any dust filters (One could be added but you have to dissemble half the PSU to get the fan screws out.). There is already a metal air guide next to the fan that is bent at a weird angle. I assume the angle is to also leave some airflow over the giant current shunt. I assume airflow is fine with the fan flipped. Only thing i did sort of mess up with the reverse airflow is the FET board over-temperature sensor. The sensor is placed on the heatsink that is farther from the fan. That way air passes across the first one and then the warm air goes over the heatsink where its sensing. That should make the farther heatsink get hotter, but now that the airflow is in reverse that heatsink is getting completely fresh air. I will probably never run it anywhere close to full power for any length of time anyway, i just wanted a PSU that does more than 3 Amps so i got this thing to really make sure i don't need a bigger PSU.
I also have a similar 4 channel version of it in the mail since i found a cheep one. But dave might have torn down one of those so its probably not worth doing a teardown myself (Apart from perhaps giving it the same fan mod).
LOL! Compared to a 3A supply, yeah, no shortage of current with this one.