[Nebby]

Quote:

I just ran an example of the B22 waterblock system. Two B22 boards share one waterblock in my design. Each MOSFET is rated for 175deg F according to the data sheet. They are attached to a 3/16" angle beam attached to a 3/16" plate.
 
According to amb's site, on a typical setup the outboard amplifier MOSFETS put out 8V*0.16A and the inboard cascode MOSFETS put out 22V*0.16A (correct me if I'm wrong, Ti). This puts 4 whole B22 boards at 38.4W. That sounds about right given its 100W power supply. For simulation purposes, let's double that for a factor of safety. That's 2.56W for each outboard MOSFET and 7.04W for each inboard MOSFET.
 
At 100% load my 125W CPU cooled by a Corsair H50 (single radiator cooler) shows 133deg F at an ambient temperature of 75deg F. Let's assume somehow we'll need to also remove 125W from the entire system and that the waterblocks are at this same temperature. When a steady-state simulation is run, the maximum indicated MOSFET temperature is 142deg F, safely below the manufacturer posed limit.
 
Take note that this temperature value is for a hypothetical extreme example. In actual use, much less power will need to be dissipated and the initial waterblock temperature will be much lower. The MOSFET temperatures will probably hover around 120deg F at most.
 
Logically speaking, if those little passive bolt-on heatsinks can adequately cool the MOSFETs, an active water loop should do the job much better.

 
I'm not too worried about the absolute temperature value as I am about the relative temperature values. Some transistor devices change characteristics depending on the temperature they are operating at so if you have mismatched temperatures it can cause issues. I can't say whether or not it's a factor with the beta22, but I think it's at least worth keeping in mind.
 
For your simulation did you factor in the thermal resistances of each thermal junction? Did you also factor in the resistance of your electrical isolator between the mosfet and the aluminum angle?

 

[haveblue]

Quote:

 
I'm not too worried about the absolute temperature value as I am about the relative temperature values. Some transistor devices change characteristics depending on the temperature they are operating at so if you have mismatched temperatures it can cause issues. I can't say whether or not it's a factor with the beta22, but I think it's at least worth keeping in mind.
 
For your simulation did you factor in the thermal resistances of each thermal junction? Did you also factor in the resistance of your electrical isolator between the mosfet and the aluminum angle?

The default recommended air-cooling heatsinking solution calls for each mosfet to be attached to the exact same aluminum heatsink. That means the inner mosfets would get hotter than the outer ones, which is apparently fine according to the designer.
 
No, I did not assume those things. Even if I did, it would not drastically change the outcome. I simplified the simulation by assuming everything was homogeneous and connected. I would have to put all 100W into one board (10x the normal estimated power output) to get the mosfets up to 175deg F. They simply don't put out enough power to remotely tax the cooling system. Thanks for the concern though.

 

[Uncle Erik]

Quote:

Uncle Erik, would you mind sharing those unusual case designs?

I fully relate with your sig, by the way.
 

Maybe not that unconventional, but I got about 90% of the way through building a tube amp with a 3/8" aluminum plate as the top with transformers, tubes and controls up top, and the guts underneath.  I actually got through most of it with handtools, but ran into problems drilling holes in the front/back pieces.  I have holes in the top and bottom plates, but was not able to drill holes through the front/back pieces to secure it, as well as a hole in the front for the headphone jack.  The front/back pieces are 3/4" thick, so they take some work.  Also, there's another 3/4" strip inside that separates the power supply from the amp circuit.  I'll also need to drill holes to mount the individual solder lugs; it will be entirely point-to-point inside.  Here's a pic:
 

 
Fortunately, I have space for a workshop now and we will start putting it together once we have the house done.  That seems to be getting dragged out a bit, but I'm hoping by the end of the year.  What I really need is a drill press to put the holes in - I broke several drill bits going by hand.  But the bulk of the tough stuff is done.  Once I get the holes drilled, I'll sand the aluminum smooth and polish it.
 
Thing is, I had probably 60-70 hours into making that case.  The original design ended up unfeasible part way in and I had to make a few compromises to get it to work.  It will work once I get my hands on a drill press.
 
Even fairly straightforward stuff like this burns hours and sent me back to the hardware store several times.  Also, I spent a lot of time sketching out the design and measuring parts before I started building it.  Then, when it all seemed to work together, I used a square, scribe, center punch, and a thin permanent marker to set up everything for drilling and cutting.
 
Even though I didn't screw that up, I made several mistakes.  First, the plates on the top were supposed to be underneath the main plate.  I was going to cut the holes, then file them square.  That proved very tricky, so I moved the plates up top.  Then the plates weren't thick enough to hold up well.  The thicker plates were then too thick to cut with my chassis punches, so I drilled them and spent hours with a file to get them acceptably round.  Another pain was mounting the transformers.  Using the nuts I wanted to use left little room underneath to get in screws to mount them to the top plate.  I hadn't thought of that before I started work.
 
I'll probably run into some unforeseen conundrums with wiring and part placement inside, but am not too worried about that.  I've restored a bunch of old radios, so I'm fairly good at stuffing parts point-to-point and getting them to work.
 
Another thing you should plan for is servicing.  One nice thing about this is that there's a bottom plate that will come off with about eight screws.  You'll be able to get to everything inside pretty fast.  If you have components stacked on components, you might have to go several hours into disassembly to get to a problem.  Troubleshooting will be tough, as well, if you have to take a lot of it apart to get to the test points or pull something out.
 
Anyhow, you run into lots of unforeseen problems when you start building stuff from scratch.  I really, really thought I had it all worked out before I started building.  I was wrong.  It took time and extra money to fix it, and it has been mothballed since around July 2009.  But I will finish it.  I like the design a lot, I want to have it in my rig and I'd love to take it to a few meets.
 
By the way, this is the Eric Barbour Brute Force amp, with the vacuum tube regulators.  I also have all the parts and aluminum to build the Ciuffoli SESS with 417A/5842 tubes, a point-to-point Bijou, Pete Millett's ECC99 SRPP, and an Electra-Print Ultrapath preamp.  There are several other projects I don't have all the parts for yet (including a Beta22), but these four are next in line.  I'm planning to build them in more or less the same style.

 

[Beefy]

This is either going to be pure awesome, or a total poochscrew. I am interested in either outcome. Subscribed.
 
Oh, and FWIW, you should seriously consider getting your transformers from SumR. He does custom potted/shielded transformers at a very good price. While I would not rely on them as a complete replacement for distance, it will certainly help a great deal.

 

[nikongod]

 
Quote:

 

You must finish this amp. There is too much awesome in that chassis to let it sit.
 
It looks a bit like a Frank Lloyd Wright prairie house style house. I like it a lot.

 

[haveblue]

Quote:

Originally Posted by Uncle Erik

That does indeed look really awesome. I have to ask though, why the use of such thick material? Is it solely for added weight?
 
Quote:

This is either going to be pure awesome, or a total poochscrew. I am interested in either outcome. Subscribed.
 
Oh, and FWIW, you should seriously consider getting your transformers from SumR. He does custom potted/shielded transformers at a very good price. While I would not rely on them as a complete replacement for distance, it will certainly help a great deal.

Thanks for that recommendation.
 
So apparently Solidworks does wiring/cables as well. I hope this will get rid of any unforeseen issues as far as routing and clearance is concerned. It'll also let me know exactly how long to cut the wires/tubes to get the desired and matching curvature, and how much to buy.

 

[qusp]

Quote:

This is either going to be pure awesome, or a total poochscrew. I am interested in either outcome. Subscribed.
 
Oh, and FWIW, you should seriously consider getting your transformers from SumR. He does custom potted/shielded transformers at a very good price. While I would not rely on them as a complete replacement for distance, it will certainly help a great deal.

indeed another vote for Richard Sumner here too, does fantastic work that compares to the big boys at much better prices, great guy to deal with too.
 
agreed, as much as I truly do hope this comes off for the OP, I will be equally enthralled by a spectacular fail, tis only human nature

 

[yellowjeep]

Bump for updates

 

[haveblue]

Just got my first full-time grown-up job, so in addition to grad classes I've been quite busy recently. Also, I'm still waiting for our school's department to either fix or get a new laser cutter. They have to get it going for their other projects by the spring. Then I'll decide whether to use polycarbonate or acrylic based on how well it cuts. Polycarbonate if possible.
 
Next semester will be less busy. I should be able to crank out the design of the whole structure within a week. Also waiting on the money
 
So lets just say that this project has been put on hold until the Spring.
 

 

[Avro_Arrow]

Welcome to the working world...
 
 
Quote:

Just got my first full-time grown-up job, so in addition to grad classes I've been quite busy recently. Also, I'm still waiting for our school's department to either fix or get a new laser cutter. They have to get it going for their other projects by the spring. Then I'll decide whether to use polycarbonate or acrylic based on how well it cuts. Polycarbonate if possible.
 
Next semester will be less busy. I should be able to crank out the design of the whole structure within a week. Also waiting on the money
 
So lets just say that this project has been put on hold until the Spring.