I considered soldering the devices to a larger piece of metal, but the thought of having to service such a thing was enough to give me nightmares. These days, people don't repair, they replace. A device that's been soldered down just goes in the trash can. I'm a little old-fashioned--not to mention hating to waste money--so I'd rather fix something if it's at all possible.
Grey
Grey
I usually take this point of view too. In other words I'm painfully cheap and have tons of used parts in boxes and bags all over the place.
However, these parts are only 94 cents each. The copper plate Variac is talking about is probably more expensive.
I think that soldering the transistors to a copper plate is a good idea. Onlinemetals.com has lots of copper sheet stock.
One idea would be to hack saw two cuts about .5" deep about .5" apart on one edge of the plate and solder the transistor to the resulting tab. It would make the soldering easier but probably not affect the cooling properties of the plate at the lower operating temperature.
Regards,
Graeme
However, these parts are only 94 cents each. The copper plate Variac is talking about is probably more expensive.
I think that soldering the transistors to a copper plate is a good idea. Onlinemetals.com has lots of copper sheet stock.
One idea would be to hack saw two cuts about .5" deep about .5" apart on one edge of the plate and solder the transistor to the resulting tab. It would make the soldering easier but probably not affect the cooling properties of the plate at the lower operating temperature.
Regards,
Graeme
3/4"x3/4 or 1"x1" are 40 cents each and require no cutting, BUT they are 18ga which isn't too thick, but copper is so heat conductive even that thickness is prob OK. Also thin helps quick cooling. after solder..
Nelson says there is no problem with clamping them on a heatsink, so this is kind of a non-issue,
BUT some people prefer having only one output device rather than many in parallel, so this is a way to get more power out of one device.
Nelson says there is no problem with clamping them on a heatsink, so this is kind of a non-issue,
BUT some people prefer having only one output device rather than many in parallel, so this is a way to get more power out of one device.
woody said:
I bet -in some non-complicated engineer way
simple clamp with screw and sort of rest?
Hi Variac,
If you are using 1" x 1" copper plates then your heat dissipation capability will be very limited.
My understanding is that you are intending to build the JFET Zen in the PassDIY article (more or lass) which means that the JFET is dissipating about 7 watts of heat. This is small compared to some of the other Zen amp applications but it is still a serious amount of heat.
I have used aluminum plates for heat sinks for several projects in the past. I have never used copper. If I was going to try and get rid of 7 watts of heat using an aluminum plate I would start experimenting with one 6" x 6" and .1" to .125" thick based on past experience. It needs to be mounted vertically and the transistor should be mounted in from the edge as much as practically possible.
Alternatively you could use one of the bigger Aavid TO-220 standalone sinks and using the mounting hole for a spring clip. You'll still have a tough time keeping the temp rise under 30C above ambient.
You may end up just going back to mounting all the power transistors (somehow) on a common large sink.
Hope this helps.
Graeme
If you are using 1" x 1" copper plates then your heat dissipation capability will be very limited.
My understanding is that you are intending to build the JFET Zen in the PassDIY article (more or lass) which means that the JFET is dissipating about 7 watts of heat. This is small compared to some of the other Zen amp applications but it is still a serious amount of heat.
I have used aluminum plates for heat sinks for several projects in the past. I have never used copper. If I was going to try and get rid of 7 watts of heat using an aluminum plate I would start experimenting with one 6" x 6" and .1" to .125" thick based on past experience. It needs to be mounted vertically and the transistor should be mounted in from the edge as much as practically possible.
Alternatively you could use one of the bigger Aavid TO-220 standalone sinks and using the mounting hole for a spring clip. You'll still have a tough time keeping the temp rise under 30C above ambient.
You may end up just going back to mounting all the power transistors (somehow) on a common large sink.
Hope this helps.
Graeme
gl
The idea was to spread the heat with the copper plate so that the copper plate can then can be attached to the aluminum heatsink. My thought was that Nelson recommends not using an insulating layer such as mica or ceramic or silicone sheet, but just attaching the device directly to the heatsink using the anodized layer of the heat sink as electrical iinsulation. Grey points out a similar approach, but use a separate heatsink for each device, and electrically isolate the entire heatsink.
My question was whether one could mount the Jfet directly to the small copper plate with solder (excellent heat transfer) then, since the plate would have about 6 times the area of the jfet, mount the plate to the heatsink with an insulating layer, which should work fine since it has a lot more area to transfer the heat through the insulator to the heatsink.
Anyway, it was just a question. I am probably just going to mount the Jfet directly to a small separarate heatsink, electrically isolated from the chassis. This seems the safest for the moment, and should get rid of the heat fine.
Thanks for the concern,
Variac
The idea was to spread the heat with the copper plate so that the copper plate can then can be attached to the aluminum heatsink. My thought was that Nelson recommends not using an insulating layer such as mica or ceramic or silicone sheet, but just attaching the device directly to the heatsink using the anodized layer of the heat sink as electrical iinsulation. Grey points out a similar approach, but use a separate heatsink for each device, and electrically isolate the entire heatsink.
My question was whether one could mount the Jfet directly to the small copper plate with solder (excellent heat transfer) then, since the plate would have about 6 times the area of the jfet, mount the plate to the heatsink with an insulating layer, which should work fine since it has a lot more area to transfer the heat through the insulator to the heatsink.
Anyway, it was just a question. I am probably just going to mount the Jfet directly to a small separarate heatsink, electrically isolated from the chassis. This seems the safest for the moment, and should get rid of the heat fine.
Thanks for the concern,
Variac
Hi Variac,
I guess I had missed the point that you were using the copper plate as a heat spreader. Sounds good. My major point was to say out that 7W was a non-trivial amount of heat to get rid of.
Regards,
Graeme
I guess I had missed the point that you were using the copper plate as a heat spreader. Sounds good. My major point was to say out that 7W was a non-trivial amount of heat to get rid of.
Regards,
Graeme
I don't want to be off thread... I want to make one!!! Let me plant a seed... I once worked with a product testing application where we had to use a Bismouth solder (might check my spelling there) on all the associated test hardware... I beleive this requirement (customer based) was to protect the integrity of the very high $$$ boards we were soldering. Bismouth is a lower temp solder. This product was also some of the highest speed ECL logic available. Very sensitive to further "processing" by over temp conditions... Just a thought that may solve that time/temp delema...
FYI we here in the USA have to hunt fowl with non Lead shot in our shotguns (HazMat). We either use steel or Bismouth shot...
So Grey, guess you did'nt like the Grey Labs thing??? Sounded good to me. You should be honored to be there next to XXXXLabs... I did not really like the flg thing when it started but I got used to it ( Found in Lee's Garage)...😀 😀 😀
FYI we here in the USA have to hunt fowl with non Lead shot in our shotguns (HazMat). We either use steel or Bismouth shot...
So Grey, guess you did'nt like the Grey Labs thing??? Sounded good to me. You should be honored to be there next to XXXXLabs... I did not really like the flg thing when it started but I got used to it ( Found in Lee's Garage)...😀 😀 😀
Didn't like? No, it just confused me. Setting aside all the usual confusions and demands on my time, there have been a lot of things going on at work and the Lovoltech buy and the author thing and Christmas and so forth.
Interestingly enough, I have four rooms in my basement. The one where I do electronics I call the Lab, with tongue firmly in cheek.
I make no claims to being in Nelson's league. He does this stuff all day, every day. I just play at it.
Grey
Interestingly enough, I have four rooms in my basement. The one where I do electronics I call the Lab, with tongue firmly in cheek.
I make no claims to being in Nelson's league. He does this stuff all day, every day. I just play at it.
Grey
I'm going on a 2 week vacation, so you guys keep the discussion going... I'll be back!
Maybe nelson will have a new version by then! (keep the bulbs as an option-please!!)
Maybe nelson will have a new version by then! (keep the bulbs as an option-please!!)
Variac said:
Sorry, the bulbs are the first thing to go. I do have a nice
Zalman "mushroom" heat sink that fits where the bulbs went.
😎
Anyone who has Zen Ver 4 boards can easily configure the boards for the Pass JFET amp with a few cut traces and jumpers. You would get a regulated supply and constant current source.
I'm back, and I plan to keep the bulbs.
I have to agree, it shouldn't be any surprise that the first improvement is a CCS. However I will console myself that Nelson seems to still like his enlightened Zen and some really like the SOZ so I'm sure it will sound great! Just will use a bit of power.
I have to agree, it shouldn't be any surprise that the first improvement is a CCS. However I will console myself that Nelson seems to still like his enlightened Zen and some really like the SOZ so I'm sure it will sound great! Just will use a bit of power.
Variac said:
While waiting for the JFets... I found nice halogen sticks for the amp .
They are 1500W 230Volt , 2Amps @ 26 volts , 13 ohm .
The stick-holders are pratical to use and easy to mount parts .
Good Year to all.
Stefano,
I'm sure people would love to see photos of the halogen sticks and holders. I have generally cleaned up the complications of returning from vacation and hope to quickly finish the amp.
Mark
I'm sure people would love to see photos of the halogen sticks and holders. I have generally cleaned up the complications of returning from vacation and hope to quickly finish the amp.
Mark
Given that the resistance of the filament in a light bulb will increase along with the temperature, it occured to me that you're introducing a distortion mechanism by having an inconstant load.
Grey
Grey
Some Years ago, after reading the Zenlite paper, I tested some bulbs myself. N.P. did not clearly state any particular conditions, but he said he tested many bulbs also. All my tests showed a variable R vs. V or I in the expected range of use. I was only doing static testing(turning knobs with my hand). I eventually assumed that the temperature lag was entering into the results. I never actually tryed an AC modulated around 2A or anything. It is very possible that N.P., in all his wisdom, just left that part for us to figure out.
For instance, I hate monitors or lights that dont go any faster than 60hz unless there persistance is long enough that I dont see the blinking. Regular incondecent bulbs rarely blink to my perception although, they obviously have a large signal 60hz on them😀 😀 😀
For instance, I hate monitors or lights that dont go any faster than 60hz unless there persistance is long enough that I dont see the blinking. Regular incondecent bulbs rarely blink to my perception although, they obviously have a large signal 60hz on them😀 😀 😀
- Status
- Not open for further replies.