In article <CANJYbi9Og4EaVcY-63Tbk5uJLmRuZcrrNa6oeyOnzSrc9Mmcng at mail.gmail.com>, Wolfgang Eichberger <oe5ewl at gmail.com> writes: Speaking of wiring up test load resistors for a PSU, just exactly how do you go about deciding what resistance and what power rating to use for a test load? -- "The Direct3D Graphics Pipeline" -- DirectX 9 version available for download <http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/> Legalize Adulthood! <http://legalizeadulthood.wordpress.com>

On 1/24/2012 5:33 PM, David Riley wrote: How would you test a AC load, say for Disc Drive for example with a power on sequence? Come Now, all speakers I can think of is reactive. What your ears are rated at is a better rating.

Richard wrote: For a basic test, where we just want to verify that the PSU works properly near its maximum rated load: rated voltage = E rated current = I resistor value = R = E/I resistor wattage = P = E*I = E^2/R For instance, for a 5.2V 35A power supply, R = 5.2/35 = 0.149 ohms. We aren't going to find an actual 0.149 ohm resistor, but rounding up a bit we can find 0.15 ohm. With that, I = E/R = 5.2/0.15 = 34.67A, near the maximum rated current, and P = (5.2^2)/0.15 = 180.27W. Note that we have to bolt that resistor to a good heatsink, and provide force air cooling, to prevent it from burning up. A 0.15 ohm 200W resistor is roughly 2" diameter by 3" long, which is not much volume (or surface area) for 181W of heat. We can buy a chassis-mount 0.15 ohm 200W resistor from Mouser Electronics, but it costs $46, and that resistor probably won't be of much use to us for anything other than testing that particular power supply. We might be better off buying more resistors of a more common value and paralleling them, or buying a variable resistor. With variable resistors (rheostats) we have to be careful, because the power dissipation rating is at the maximum resistance, and for lower resistance we have to linearly derate the power. For instance, we could try using a 225W 1 ohm variable resistor, but when we set the tap at 0.15 ohms, we now have to derate the power to only 33.75W. That won't meet our requirements. Suppose instead we parallel 33 5 ohm resistors. That gives us an overall resistance of 0.152 ohms, and power dissipation of 5.2W each. We can use resistors that cost us $0.93 each. In total, that is $30.69. We haven't saved much money over the single resistor, and now we have to expend more effort in wiring the resistors up to form the load, but it's far more likely that we'll be able to reuse those resistors for a different power supply load in the future. To do "serious" power supply testing, we would need to measure line regulation and load regulation. For load regulation measurement, we can put a solid-state relay in series with some of our parallel resistors, so that we can near instantaneously change the load current (load step), and watch the supply voltage change on an oscilliscope. Lyle Bickley and I built the test loads for the DEC 728 power supplies of the PDP-1. We did not test line regulation and load regulation for step response; we just tested the power supply under four different load conditions, and measured the DC voltage, RMS ripple, and peak-to-peak ripple of each output under each of the four load conditions, with a variac on the input at 105V, 115V, and 125V. Testing the 728 is complicated by the fact that it is a dual output supply, providing +10V and -15V. There is a maximum current rating for each output, but there is also a maximum overall power, such that one cannot draw the maximum current from both outputs simultaneous. We built the load box with a rotary switch to select load combinations of: +10V at 0.2A, -15V at 0A +10V at 7.5A, -15V at 0A +10V at 0.2A, -15V at 8.5A +10V at 2.0A, -15V at 7.0A Eric

I start by knowing/estimating a suitable load current. If the power supply ratings are given (e.g in a service manual),m then I ill pick somethign that's around 10%-50% of the maximum load (if there's a minimum load figure given, I'll make sure the current I use is above that, of course). If I don;t have nay details on the PSU, I have to estimate the current. In the case of a lienar PSU, that's quite easy -- eitehr from the types of 3-terminal regualtors used, or from the design of the over-current trip (if it uses a 723, say). For a switcher it gets harder, but you may be able to make some sourt of guess from the circuitry it powers. Then calcualte the resistance in the obvious way (R=V/I where V = the voltage of that PSU output, I = the current you'be just guestimated). And the power dissipated in the load is of cource V*I (It's DC, so there's no power factor to worry about). Often car bulbs amke a suitable loard. It's worth rememebriogn that a filament lamp has an approxximately constant-current (not constant-resistance) characteristic. So a 6V 24W bulb will draw 4A from a 6V supply. I'tll also draw close to 4A from a 5V supply. -tony

Have you tried to completley dismantle the RC50? It's a bit strange. The outer can comes off trivially, then after htant I think you remvoe the sideses, the seek/interface PCB fro,m the back (stepper motor controlelr). Then free the front casting and take off the disk clamp arms (there are C-shapped sring rings whichleft thes when you open the doors. After that the R/W and spindle motor PCBs come off trivially. So does the belt, the spidnel motor and the hubs. But don't touch the stepper motor or heads. The RX50 seems ot haev been desinged so it's almost impossible to alight -- you can't get to the stepper mounting screws when the seek/interface PCB is fitted, conversely you need this fitted to runs the drive. I have a LART resereved fro the 'genius' who came up with that one... -tony

I think I managed with some extension cables between the seek/interface PCB and the other PCBs (and the stepper motor, sensors, etc) when I did the one in my Rainbow. Still a right pain... I've seen much worse. Ones built on a plastic chassis for example. When I got mine working, it seemed pretty reliable actually. -tony

Hi all, I was quite busy this week - so I had not much time to play with hardware.... .) no, I did not disassemble the whole RX50. Only a good cleaning was necessary and that sort of stuff.... .) what I did was checking the backplane against the manualy I found and it seems to be fully functional; the frontpanel (okay, it's only 3 switches and 3 lights) too. .) the PSU is still object for work. @Rob: I did hear about BOS but never have seen any material survived. If you ever find your disks, please think of me. -- Wolfgang Eichberger - OE5EWL Operating System Collector Blog: 5ewl.blogspot.com Homepage: www.eichberger.org 2012/1/26 Rob <robert at irrelevant.com>