New test load

[Meatball]

The entire back side of the PCB is a copper layer, with vias holes that go to the resistors.
With just that bar, I was able to give it 3.3A for 20 seconds before i had to take my thumb off the top of the diodes (the driver was another story...i could smell that poor regulator)
I'd like to see 1 minute run times without the readings straying too far off.

DTR has all the proper equipment to accurately test it, so i'm mostly relying on him for the results.

I've got 1/2" bar coming which will cover from the white line to the edge (twice as wide as the current one)
I've also got small finned copper heat sinks coming. They are the Chinese ones made for RAM chips.
Not sure how i'll incorporate them until they arrive. I thought about lining them up and soldering them together into a strip to mount on the board.

I didn't get the test load mailed today, I got caught up with the lender over our new house, as well as working today so I can have tomorrow off. We're closing on our new house tomorrow, so i'm going to drop it off at the post office on our way to the title company.

Prices is really start to go up on these. I want a nice quality test load that a person only needs to buy once and it will last them for as long as they're into lasers, and be less than $30 retail. Though when I order in bulk it will drop the price by a few dollars each...

I think you make a good point moh, unless you're looking at test longgggg term stability of a high current driver, the resistors on this board ought to be just fine.

It takes time to heat resistors this size. All it takes is 5 seconds to know where your driver current is landing. If you need 3 minutes to tell if your driver is in the correct range, than you can simply adjust the dial without it necessarily being on. Aim and shoot until you get the voltage you want. If you need to sink 3A for ten minutes for testing, yeah maybe you could actively cool something this small.

 

[Mohrenberg]

DTR should get the test load in the mail today. I've been busy with the new house, and will be gone most of this coming week as the actual move is coming. Also have carpeting tiling, fencing and some painting left to do....

 

[Benm]

I am somewhat confused. If yo already have that big slab of copper there, why not just bolt a finned heatsink onto that so that it can sustain continous operation at full power? It seems like a very cheap addition that turns something impractical into something more or less foolproof.

 

[Bionic-Badger]

Well, for those particular diode packages the way he has it is appropriate. Perhaps spreading those diodes around with more copper between would've been a good idea.

The real problem is that for such high current loads a different package like a TO-220 would've probably been better. Then the diodes could be bolted to an upright finned heatsink attached to the PCB substrate. As is, a finned heatsink would need to be attached to the bottom of the PCB like that copper bar, and then bent upwards, or protrude downwards. Maybe a successive iteration will improve on the heat dissipation. I don't know if the added cost is really worth the effort though.

 

[benmwv]

The problem is finding a finned heatsink that will fit it. The heatsink cant go past that white line because of the pins sticking out. Maybe some kind of extruded aluminum could work.

 

[BShanahan14rulz]

Or you could mount it to the edge of the heatsink to avoid shorting those pins

or you could mount the heatsink to the copper heat spreader

I say keep the copper bar, and put a warning that if you want to run the test load for more than 30s (or whatever time that testing shows the copper alone will suffice for max current), you will need to source and install your own heatsink to the copper heat spreader.

Next, people will be expecting drivers to come pre-installed to the test load... :tired:

Folks, don't forget, dude needs to be able to recoup his investment, so he can't put every single feature in from the start, would get too expensive!

 

[gillza]

Select-able test-load with multiple pins and two types of resistors to address the voltage over the resistor drop issue!!! Now that I'd definitely buy!!!

CongratZ Benmw and Mohrenberg!

 

[Mohrenberg]

Video from DTR.

I have to say it works pretty well. Only about a 200mW drift after 3 minutes. I found J3 is pretty much a perfect match for M140's and 9mm's. @ 4.5V you are around 1.8A which is perfect and @ 2.2A you are a little over 5V so that matches both diode for the common currents they are set to.

Here is a video. Power is applied in constant voltage mode to show the voltage drop at particular currents.

 

[Mohrenberg]

I'd like to add single post screw terminals to the DMM pads, so people can make leads to their DMM, or just jam their dmm probes right in the hole.

Aside from that I think we just have to move some of the labels around on the board and it'll be ready.

DTR's video was done with the smaller heat sink as well. The retail will have 1/2" wide copper bar instead of 1/4"

 

[Elektrotechniker]

I am in for one of these when they will be available I really like the look of it

 

[DTR]

Ok guys here are some voltage drop tests with each of the jumper settings.

I have not actually seen a test load yet that can simulate a 405 diode perfectly as you need a lot of diodes on there. The J5 @ 2A is dropping 6V but @ 450mA which is where I usually set my 405's the drop is only 4.5V.

 

[benmwv]

DTR, remove that jumper completely and you will have one more diode added in. That should put you up around 5.3v.

 

[DTR]

That would be better. I will get a test of how that turns out but I would love to see one that actually lands about 6V @ 450mA so I can properly test the dropout on linear or buck drivers or see that a new boost driver I am testing is not going to have trouble getting to the voltage requirements of the diode.

Also let me be clear this is not a flaw in this test load as none others I have tried can do it but it would be nice to see one have that capability. Either way this is a great load and it is very stable with the higher currents for the 445's. Can't tell you when the 445 first came out how many test loads I fried from overheating and the pain of trying to set a pot when the readings go crazy after a few seconds from the heat.:beer:

 

[Blord]

I bought some copper heatsink for the testload. 8 pieces for $4.50 shipped, http://www.ebay.ca/itm/ws/eBayISAPI.dll?ViewItem&item=330874686360&ssPageName=ADME:L:OC:BE:3160
They were designed for cooling the video RAM chips and have adhesive tape on them.
Four of them in a row fits perfectly on the transistors.
The loadload can also put in the Amp mode of the multimeter. Just connect the yellow encircled points and you are omitting the resistors completely.
I have measured both ways and the results were within few percent.

 

[Speedy78]

I got some of the copper from him and picked up the hardware at the local home improvement store.