I know that there is about a billion questions on the internet that asks "what electronic load to buy" but funny enough have I not found anyone asking: "What model do have least self-noise"?
I have read all sort of articles, forums and whatever but have only found a "buy this or this" for your price target. Or is a DIY better then a buy.
I am after an electronic load that I can test power supply's with, one that do not make more noise then the unit I try to test and one that can be used to test some of the best audio amplifier psu's with.
My price is not set at the moment, I'd rather save up a half year more and get what I need then have to live with something that do not quiet work.
Please if you have any idea on what to choose, low noise is a must, great user interface would be nice and high power would be great.
I have read all sort of articles, forums and whatever but have only found a "buy this or this" for your price target. Or is a DIY better then a buy.
I am after an electronic load that I can test power supply's with, one that do not make more noise then the unit I try to test and one that can be used to test some of the best audio amplifier psu's with.
My price is not set at the moment, I'd rather save up a half year more and get what I need then have to live with something that do not quiet work.
Please if you have any idea on what to choose, low noise is a must, great user interface would be nice and high power would be great.
I'd use a passive resistive load for critical testing, no noise there. Carborundum high power types
work well, and have been long used in power RF for their low inductance.
work well, and have been long used in power RF for their low inductance.
Last edited:
Yes that is noiseless🙂 But it is wary hard to do, i.e. pulse testing.I'd use a passive resistive load for critical testing, no noise there. Carborundum high power types
work well, and have been long used in power RF for their low inductance.
What I like to get is a reliable instrument, one where I know it is precise, maybe even a high slew rate.
A passive load will have Johnson noise equal to sqrt(4kTRB)
An active load is probably going to be noisier, but none will be noiseless. Absorbing power is a lossy process, so thermodynamics says it has to be noisy.
An active load is probably going to be noisier, but none will be noiseless. Absorbing power is a lossy process, so thermodynamics says it has to be noisy.
One thing I've wondered is whether you could use a lab supply set to constant current mode to load the supply and therefore act like an active load? Supplies like the HP 6627A seem to be finely programmable and equipped with readback, so while they can't be programmed directly as a load, with a little math, you could come up with the current required to behave like an active load. These sort of supplies also seem to be plentiful and less expensive than an active load as well.
Yes you are right, that's also why I wrote it in quote marks🙂A passive load will have Johnson noise equal to sqrt(4kTRB)
An active load is probably going to be noisier, but none will be noiseless. Absorbing power is a lossy process, so thermodynamics says it has to be noisy.
Dave from EEVblog did have some real problem until he found out that his BK precision load, did make all the noise, and did make it almost useless to test his cheap Rigol power supply. I would like to avoid the same problem, but still have the features and quality of a commercial model.
Fluke supplies (don't remember the model N°) do that: as long as the load is passive, they behave just like any other, but if the load is a generator and exceeds the set voltage, they switch to a current sink, with a value identical to the current limit presetOne thing I've wondered is whether you could use a lab supply set to constant current mode to load the supply and therefore act like an active load? Supplies like the HP 6627A seem to be finely programmable and equipped with readback, so while they can't be programmed directly as a load, with a little math, you could come up with the current required to behave like an active load. These sort of supplies also seem to be plentiful and less expensive than an active load as well.
Constant Current (CC) mode of power supply is usually really noisy, at least an order or two noisier than constant voltage mode.One thing I've wondered is whether you could use a lab supply set to constant current mode to load the supply and therefore act like an active load? .
Electronic load usually not very noisy (?) because of integrator type schematics of el. load. And the slower the load - the lower frequency integrator is.
Last edited:
The idea of using a power supply is interesting and fine if you just need an adhoc load, but what I try to find is a load that can torture a power supply and be used to test if it is well build within spec. I know that Kikusui do make some of the best, but they are also way up in price, so maybe a compromise between great performance and price🙂
If TS is looking for diy device, than here it is:
A Closed-Loop, Wideband, 100A Active Load
And the same.
If he searches something manufactured ready-made - I'm not experienced in.
A Closed-Loop, Wideband, 100A Active Load
And the same.
If he searches something manufactured ready-made - I'm not experienced in.
Last edited:
Can any of you please tell me if these specifications is realistic?
Lets say I want to test a power supply up to a 200W 8 Ohm. It can also be 800W 2 Ohm.
sqr(200x8) = 40V 5A or sqr(800x2) = 40V 20A.
Add some headroom and you get 50V 25A or 1,250W load
Lets say I want to test a power supply up to a 200W 8 Ohm. It can also be 800W 2 Ohm.
sqr(200x8) = 40V 5A or sqr(800x2) = 40V 20A.
Add some headroom and you get 50V 25A or 1,250W load
The voltage is a little low - 200W RMS into 8 ohms is 40V RMS, multiplied by sqrt(2) yields 56.5V peak. Add a few volts for output stage headroom and the supply for a 200W 8Ω amplifier will be closer to 60 or 65V bipolar. The current required for 800W RMS into 2Ω would be 20A, but again, that is the RMS value, so the peak assuming an 800W sine wave spec would have to be 28.2A, requiring a 30A spec with probably a little extra headroom above that.
Oh, thanks a lot, so the formula should be sqrt(2) x sqrt(WxR) + Hr = Vp/p ?The voltage is a little low - 200W RMS into 8 ohms is 40V RMS, multiplied by sqrt(2) yields 56.5V peak. Add a few volts for output stage headroom and the supply for a 200W 8Ω amplifier will be closer to 60 or 65V bipolar. The current required for 800W RMS into 2Ω would be 20A, but again, that is the RMS value, so the peak assuming an 800W sine wave spec would have to be 28.2A, requiring a 30A spec with probably a little extra headroom above that.
WOW 1800VA sounds like a lot!! 🙂
I'm not sure that we need a noiseless dc load for testing that power supply, if it's only ones. Maybe just powerful resistance, or something like a spiral nichrome heating element from electrical heater.
Last edited:
Hmm would it be possible connect several loads in series or parallel?
The outputs are floating, but check the voltage isolation specs.
Thanks! 🙂The outputs are floating, but check the voltage isolation specs.
So as long the ground are floating and total V/A does not exceed isolation limit, all should be fine, do the same apply for power supply's?
What I mean is get two or more units of the same type and model, and then just connect them in either series or parallel?
Generally lab supplies have floating outputs, but verify this before use. Also check the voltage isolation.
While typically two supplies can be connected in series, don't assume that you can float the supply
from ground by more than its maximum output voltage unless the specs say so.
While typically two supplies can be connected in series, don't assume that you can float the supply
from ground by more than its maximum output voltage unless the specs say so.
Last edited:
- Home
- Design & Build
- Equipment & Tools
- Buy - "Noiseless" Electronic load