[dnewman]

Quote:

Originally Posted by Nerull /img/forum/go_quote.gif The bill of materials is attached below.

While many of us have lots of spare heat sinks lying about, you may
for the sake of completeness wish to add them where needed to the
parts list.

Cheers,
Dan

 

[ATAT]

Nerull can you post black and white pictures of the outline of the traces only? Im guessing I could rout this on a cnc.

 

[dnewman]

Quote:

Originally Posted by dnewman /img/forum/go_quote.gif FWIW, using a power regulator for the filament DC supply is extreme overkill. Keeping in mind that those heaters will work off of AC just fine (albeit with a shortened life expectency), most manufacturers including high-end just use a bridge rectifier, some caps, and some resistors. No regulator at all. I'd even argue that having two xformer secondaries instead of just one for the heaters in this specific design is overkill.

Seeing that the 6336 tubes can eat up to 4.75A apiece with
their heaters, I've revised my opinion of using a regulator. The
size of the capacitors you need to use to get even ~1 VDC ripple and
handle the 2 x 4.75A + 2 x 0.6A (12AX7's two heaters) is in the
range that a LT1083 becomes attractive....

Dan

 

[dnewman]

Here's a schematic for an adjustable 6.3V DC heater supply with slow start.

In the schematic
1. All resistors are 1/4 W.
2. All capacitor voltage ratings are working ratings in DC volts.
3. The diode D3 is only needed when the voltage regulator (IC3) sees
output capacitance of 1,000 uF or more. In that case a crowbar
situation can arise.
4. The total capacitance of the parallel bypass capacitors C9 - C13 may
be reduced when the input AC voltage is increased. Tradeoff is
ripple vs. regulator dropout voltage and the increased power dissipation
in the regulator when higher input voltage is presented.
5. At a ripple frequence of 120 Hz and R5=100R, C14 should be 25 uF
+/- 20%. However, using 22 uF is fine.
6. R6, R9, and R10 may all be replaced with a single 402R resistor. Use
of R6, R9, and R10 allows for output voltage adjustment between 5.7 - 6.6V.

FWIW, I have laid out and routed an entire Bamaslama PSU PCB on a 5x7" board
with all the supplies (+/- 150 VDC, +/- 15VDC, +6.3VDC), heatsinks and whatnot.
Have also routed for both 1 and 2 oz copper. (Issue being the high current in the
6.3V -- heater -- supply which needs ~100mil tracks on 2 oz copper and ~200mil
tracks on 1 oz copper.)

 

[dnewman]

There's some errors in this bill of materials. Most importantly are that the
Mouser part numbers for the 4.7V zeners are for surface mount versions.
Folks probably want an axial DO-41 package instead.

More significantly, the mouser part number for the OP27 is an incorrect
part. Not only is it surface mount, it is a part which is not electrically
compatible. It's a part which Mouser turns up as a cross reference
for an OP27. However, the cross referenced part is not suitable for
this application: it's maximum voltage ratings are significantly less than
the +/- 15VDC required by the Bamaslama. Additionally, the part isn't
pin wise compatible with an OP27 either (i.e., won't work with the
pinouts assumed on Dr. Gilmore's board).

Regards,
Dan

 

[Nerull]

I finally have the money to build this thing, and motivation to do so now. dnewman, I'm currently contemplating whether or not having a DC heater supply at all, seeing as it really doesn't affect anything based on what I've read (diyaudio.com mostly). I don't think it may be worth the time and effort to do so (and make things really hot in the case).

I will also update the BOM to accommodate for the changes made to the power supply and corrections as said by dnewman (thanks, I guess I wasn't really thinking). As for the old board layout, it wouldn't work due to the immense heat given off by the IRF644 MOSFETs and 2SA1156s (thus requiring heatsinking).

More information will come as I think of it (and come to investigate further.)

~Tom

 

[dnewman]

Nerull,

For IRF644s, consider heatsinks along the lines of Mouser #532-529902B25G
(Eagle PCB SK129@heatsinks.lbr). I'm planning on using them for simplicity for the 2SA1156s also. And, don't forget the messy heatsink
thermal compound and some 4-40 (or so) screws to attach said devices
to the heatsinks. (One of these days I'll run out of my thermal compound
and try the new generation of "gaskets" that are available.)

For the +/-15V regulators, I'm planning on using Mouser 532-592502B34
(Eagle PCB FK218). If you use your MP925 packaged 50K / 25W resistor,
then you could use the same style heatsink. However, I suspect that that
part won't need a heatsink if it's only dealing with 2W.

And yes, if you want to use 6336 tubes, then the DC heater supply becomes
even more of a monster than that which I posted above: instead of 6.2 A,
you'll be eating closer to 10.7A (4.75x2 + 0.6x2). You'd need two LT1083s!
(I figure I'll try DC on less power hungry heaters and then if I ever want
to try 6336s, I'll try them with AC and hope that it won't be too much
of an apples and oranges comparison.)

Ciao,
Dan

 

[Nerull]

Alright, would someone like to head the group buy? I don't have experience doing something like this (and don't want an angry mob at my doorstep). I don't think it'll be large scale, so it won't be too intimidating. But will someone try to head it though? (I have the DynaTube gerber layouts ready to be sent to imagineering pcbs.)

~Tom

 

[Uncle Erik]

Nerull, I think we have to get permission from the admins to do a group buy. I'm short on time these days, but I would buy some of the boards.

 

[a1rocketpilot]

If someone is willing to head up a group buy, I would certainly be in for some boards.

Aditya

 

[Nerull]

I currently have permission (got it awhile ago), but I don't exactly feel up to the responsibility of handling it (unless it comes down to me), then I guess someone has to do it

.

~Tom

 

[dnewman]

Hmmm. So much for having e-mail sent to me when this thread is
updated...

Now, I have three (extra) Bamaslama PCBs -- one is the solder mask test
board but should be functionally okay. These were made from the
Gerber's which Dr. Gilmore made available with *one* change: the
relay PCB holes and pads are reduced to only accomodate a PCB
mount relay (like the Omron in Nerull's BOM). The original Gerbers
had holes large enough to put solder tabs through. I'll gladly send all
three to Nerull: one for him to keep and two others to disperse
as he sees fit. I don't need to be reimbursed.

[As an aside, if you use Dr. Gilmore's Gerbers, note that the top and
bottom side "mask" files call for solder mask over the ENTIRE board,
pads and all (.BM and .TM files). It's the top and bottom "resist" files
which actually provide the solder mask info (.BR and .TR). I've not
used Proteus which is what Dr. Gilmore used: maybe that's just the
terminology it uses. At any rate that threw me for a bit as I'm used
to "mask" meaning "solder mask" and "resist" meaning the negative
or positive image you'd use if doing etching of the copper.]

Also, for the time being I'm going with 6.3VAC for the heater coils.
The current draw on the transformer with all those capacitors used
for a regulated 6.3 supply is murder when you contemplate using
6336 tubes. I recall current draws approaching 18A RMS
for the first few seconds as the tube heaters heated up (and hence
resistance went up, reducing current draw). This from modelling
with LTSpice.

And, finally I'm re laying out the Bamaslama board for purposes of

1. Having thermals on the ground planes,
2. Being <= 64 sq. inches (see, e.g., Imagineering's "introductory" pricing),
3. To accomodate a variety of tube socket types,
4. To fit in the Lansing enclosure I want to use, and
5. To mount the tube sockets from the solder side of the board and
then mount the board inverted (component side down), tube
sockets up. PSU on a separate board which would not have the
components upside down -- heat rises after all and no point trapping
it under a circuit board.

I'm doing all of this in Eagle PCB. If anyone wants the Eagle version of
the Bamaslama schematics, let me know. They are in 5 sheets so I
guess I'll need to collapse them to one sheet for folks with the free
version. Also, I have some "custom" packages so a slimmed down
version of my custom parts library would be needed. Ditto this for
PSU schematics.

Cheers,
Dan

 

[dnewman]

FWIW, here's a snap of my layout of the Bamaslama board. It's 8x6.5" and
intended to fit into a 2U high, 1/2rack wide, 10" deep enclosure along with
it's power supply but with the potted tran on the outside. Tubes mount on
the underside of the board and it is then inverted and hung from the top lid
of the enclosure. (I prefer to not mount the tube sockets to the chasis: a pcb
has better dampening properties.)

 

[Veniogenesis]

Hi everyone! Is there a completed/finalized BOM somewhere? Also, what is the word on a group buy of finalized PCBs?

Cheers,
Venio