[tinkerman]

Originally Posted by owdi

Yes, a blow dryer has no switching, it's 90% resistive heating and 10% reactive AC motor.
... If you make the inverter drive a switching power supply I think you'll have a harder time finding the current limits of the server PSU. ...

In another forum (EEVBlog) the electronics/power geeks have commented that a heat gun (with TRIAC switching heat control), even if it were dialed down to say 700W would cause current spikes with power beyond the considered PSU output ratings (1200~1500W). (?!)

If the above is indeed true, it would mean that current spikes could be high enough to trigger the OCP(Over Current Protection) in the PSU and turn it off, or cause other issues with the PSU. I'll be looking into this further and update if possible. References are difficult to find.

Returning to your comment that I'd "have a harder time finding the current limits of the server PSU", this is in direct contrast to the comments I encountered in the other forum. So I must ask:

1. What watt rating hair dryer(s) have you used?
2. How much current did the hair dryer actually pull from the PSU?
3. What were the "unexpected results" of your tests? (with regard to the use of inverters and hair dryers)

TRIACs are like a can of worms... interesting but I'm starting to see the attraction of hair dryers!

Thanks...

[owdi]

1. It's a small travel blow dryer. It pulled about 200 watts on low and about 800 watts on high. I say "about" because I didn't try to measure exact power of the blow dryer, I only cared about how much current the inverter pulled when powering the blow dryer, and how far PSU output voltage sagged.

2. The blow dryer was connected to the inverter, which then pulled about 55A with only that blow dryer on high. I think the modified square wave from the inverter caused it to pull a bit less than a nice sine wave from wall power.

3. My only unexpected result was the output of the HP PSU, 88A at 13.5V is nuts.

I used several DC devices to load the HP PSU:
800 watt inverter
450 watt inverter
Old class AB car amp

...and several AC devices to load the inverters:
Travel blow dryer, changed between high & low
Travel iron set for 230V (causing it to pull one quarter the power as the 115V setting)
Travel boot & glove dryer
Desk lamp with 40 watt incandescent light bulb

The above is why I don't have measurements at nice intervals. I mixed stuff together then measured the current and voltage at PSU output.

Regarding TRIACs, the thing to understand is they are just fast ON/OFF switches for AC circuits. That 2000 watt heat gun will not draw 700 watts when you adjust the output knob. It will draw 2000 watts on a 35% duty cycle for an average of 700 watts. Turning the knob only changes the duty cycle on the timing circuit that controls the TRIAC. The inverter needs to output 2000 watts, briefly, to power the heat gun at any output setting.

It's the same as modern dimmer switches, which need special LED lights to avoid flicker when dimmed. The more you dim the shorter the duty cycle (more time spent OFF). Make sense?

This is what I meant by "strange" results. The PSU and inverter both have filtering capacitors to handle current spikes, but blasting them with short 2000 watt bursts will likely yield different results than a constant 700 watt load when exploring the limits of PSU output.

[tinkerman]

Thanks for your detailed update. Very handy info there.

After quite some thought and shoppin' around, I've realized that the most cost effective, simple, and safe type of dummy load (albeit potentially a bit bulky) would be to get a bunch of 12V 50~60W halogen bulbs* + ceramic sockets (both very cheap these days at around oh $0.50 or less each). Switches (also pretty cheap) for each bulb would allow testing in 50~60W steady increments, and prevent triggering the PSU's OCP circuit.
* Apparently 12V bulbs will endure slightly higher voltages(at the cost of shorter lifespan), so I'm thinkin' I should be able to get away with say 13.5~14V for an hour or so. And there wouldn't be crazy ripple in the PSU resulting from "switching" at the load as would be the case with a TRIAC heat controlled load.

Money saved with the PSU and testing apparatus would be handy towards a genuine ICOM Next from the BMW Service Center (luckily offering at a very reasonable price, albeit have to expect to wait 4~6 weeks for delivery). Fortunately not in any rush...

Cheers~

[owdi]

Apologies for continuing to rain on your parade here...

The resistance of halogen bulbs changes with temperature. It's much lower when they are cold, then increases as they heat up. Per the second post in the below thread, at room temperature the resistance will be 10%, so those bulbs will try to pull 500 watts each before they heat up.

https://www.eevblog.com/forum/beginn...al-resistance/

To do this right you need wire wound resistors with heat sinks. The lowest cost option I could think of was using 6x 100watt 1ohm and 3x 4ohm wirewound resistors with aluminum heat sinks, then add active cooling with a fan so they could dissipate 200 watts each. All 9 wired in parallel would be .148 ohms, and would pull about 1300 watts depending on voltage. Those resistors cost about $1 each on aliexpress, or $2 on amazon.

[tinkerman]

I actually appreciate your feedback... it's been quite stimulating.

If I go the halogen bulbs route, I'll have to make an aluminum sheet metal or foil shroud perhaps with vents that can be closed and opened, then cycle through the bulbs to preheat them with the shroud closed, then gradually "re-power on" bulbs one at a time and open vents on the shroud for ventilation gradually as well. Just an idea...

I considered wire wound resistors a while back but was discouraged by the following video, which demonstrates a simpler(?) method - use of enamel coated wire in a tub of water as a very simple solution for high power loading.

[owdi]

Eeek, that video highlights the danger of cheap components. It sure looks like these cheap load resistors with aluminum heat sink can't really sustain their rated load.

The simpler method of immersing wire in water is wild, maybe it works as a quick solution, but it's nothing I would personally try.

One option is 1 foot of 14 awg nichrome wire immersed in a bottle of mineral oil. That gives the target .15ohm resistance for 1200 watts at 13.5V. The problem is it's not adjustable, the load is always the same.

Another option is galvanized fence wire, the same stuff that chain-link fences are made of. That has 10x the resistance of copper, so about 12 feet of 11 awg would give the target resistance, and could be bent into a coil that a fan could cool, without the need for mineral oil.

Neither option is as good as parallel load resistors, since those let you fine tune the load by connecting/disconnecting resistors of different sizes. You just have to size each resistor correctly for the voltage it will see. At 14.1V, 1 ohm is 200 watts, 2ohms is 100 watts, 4 ohms is 50 watts. A bank of 2 ohm 100 watt resistors would be just right.

[tinkerman]

If shopping in Aliexpress, nichrome wire 1.5mm diameter(note 14AWG = 1.63mm) x 1m(length) can be had for less than about $3 shipped!(to Korea, likely a bit more to other regions) So several lengths could be made for the power/current of interest. A liter(or more which is likely necessary) of mineral oil costs more of course. But this would indeed be a very inexpensive method to make a dummy load.
I found a handy online nichrome calculator(in metric) for determining one's own lengths based on different diameter of wire:
https://calculator.academy/nichrome-...ce-calculator/

14 x 2 ohm 100 watt resistors from Aliexpress could be had for under $14 shipped!(to Korea) As I would consider these consumables for testing a few times only, I could accept cheap resistors(branded ones like ARCOL cost like $10 EACH locally so not considering them). A good sized multi-fin aluminum heatsink/radiator would also have to be thoroughly considered for thermal dissipation capacity and price, which will not be cheap either(though I have an oversized aluminum oven pan which would work, LOL). But I definitely see this option's appeal for ease of varying the load in a compact package.

I'll have to think about these 2 attractive methods a bit more... shop around for better sources of the pricier materials.

Cheers...

[tinkerman]

I was checking out the specifications sheet(sample attached below) for the Deutronic 1200 PSU, noting things like "recovery time", "ripple and noise"* etc..

* Data Sheet indicates:
"Recovery time" : <(less than) 1ms
"Voltage ripple" : <(less than) 50mVpp(mV peak-to-peak)
"Switching spike" : <(less than) 300mVpp(mV peak-to-peak)

It would be interesting to compare such specs with that of server PSUs (to get an idea of lab quality noise levels and those good enough to get the job done).
Has anyone checked such specs (under load), or can post specs(from data sheets) for particular server PSU models?

Thanks...

Update
The following link downloads Infineon Technologies' application note paper on an "800 W Platinum® server power supply".
https://www.google.com/url?sa=t&rct=..._&opi=89978449

In the paper's "6. Test Results" (p.42/70), the following is stated:

Test: Steady-state Vout ripple
Conditions: 230 Vrms/115 Vrms/90 Vrms 50/60Hz, 10% to 100% load
Result: |∆Vout| < 120 mVpk-pk*
* More than twice the value of the DBL 1200(Deutronic's specs are relatively more restrictive)

As detailed server PSU specs from say HP and Dell are difficult to find, Infineon's example at least serves as a good reference. Hopefully, actual specs of HP server PSUs will be found...

[tinkerman]

HP does not reveal 'ripple & noise' of its server PSUs, but Delta, one of the manufacturers of HP's PSUs, indicates the spec in its brochure*:

*Delta AC-DC CRPS Front-End Power Supply
https://downloadcenter.deltaww.com/d...ath=1&hl=en-US

Output Characteristics
Ripple & Noise - 12V: <120mVp-p

It seems that "less than 120mV peak-to-peak" is sort of the standard for server PSUs, albeit there are ones with lower standards like 200mVp-p (e.g. SUPLET SAC1600).

Update
'Voltage Regulation':
DBL 1200 = <5% 100Hz (10~90% load)
vs
Delta PSU = +/-5%

"Recovery Time"/"Rise Time":
DBL 1200 = <1ms
vs
Delta PSU = <70ms(DPS-1300-AB-21 N)

The similar voltage regulation capability of server PSUs seems to be the key factor for its ability to be used during automotive ECU flash/programming.

The less restrictive yet still low ripple & noise spec and the slower "Recovery Time"/"Rise Time" of the server PSU seem to be good enough, likely due to "support" from the car's battery acting like a buffer/capacitor.

[mattanderson]

Any thoughts on using an "RV Converter" as a power supply?

e.g.:

https://www.amazon.com/PowerMax-PM4-.../dp/B01ER3LH5W

https://www.amazon.com/RecPro-Conver.../dp/B092W7XW9J

I'm a big fan of DIY projects and wiring projects, but messing around with high power + lack of protection circuits isn't my favorite time to step outside my comfort zone. Would love to just find an easy solution that doesn't break the bank.

[tinkerman]

The RecPro units' features includes "3 fixed output modes: 13.0V, 14.6V, 16.5V". It is not variable so not ideal for programming ECUs. It would be best to be able to adjust the voltage just above 14V so that even considering a slight voltage drop, a program like ISTA or ISTA-P(for E series) would operate with steady voltage above 13V and just under 14V.

If the Powermax units indeed have "Fully Adjustable* Output from 13V to 16.5V"(*if that means it is continuously variable), then the 75Amp model would be plenty to get the job done.

If you have an RV, and need such PSU anyway and the option to use it for the occasional flash/programming would be just a bonus, then well, that may be a great option!

But I'd certainly seek out a more detailed specifications sheet(you may have to contact the manufacturer or domestic distributor and ask really nicely for detailed specs) to make sure the voltage regulation, output ripple & noise, rise/recovery time etc. are not too off compared to that of the server PSUs, as these RV PSUs are not designed to power sensitive electronics like a server in 24/7 operation(even less so automotive ECUs), and therefore may not be as robust in specs or quality as server PSUs. Or better yet, if you can find an RV enthusiast that also happens to be a PSU or electronics engineer, that can provide a review, it would be great to hear about it!