Quote:



Just checked mine - nope, no bias change with volume, and I listen to 12AE6/A's all the time - they're my favorite. I'm in the minority, though. Most people agree with your tastes in Millett tubes.

Just out of curiosity, what's the rating of the caps you used for C4L/R and C5L/R?

Yay, it's working. Tonight I got my second MAX going, this one with 2SC2238/2SA968 transistors in the BJT diamond bridge and Du Mont 12FK6 tubes. (I've got some RCAs which I may use, but I wanted to try these for now.)

Because of lessons learned with the previous MAX I built this one went together a lot better. I reversed the trimmers as needed to make counter-clockwise always equal the lower value, fitted wire loop test points and filled the test pads to make biasing easier, used slow-blow fuses, and temporarily borrowed a few parts from my older MAX until I get the new case finished. All in all it's working out great.

I ended up using 25V caps for CA7L/R, which is pushing it, but I don't have a chance to get 35V parts yet. As soon as I do I'll replace them. (I watched these positions ramp up to 25V-26V then back down on powerup on my MOSFET MAX, so I know this is risky... I'll meter the BJT version later.)

I did stupidly cut my thumb trying to remove the serrated washer from a 4-40 sems machine screw, but strangely it didn't hurt and is already healing.

I'm just glad this is working, because now I can get on to fashioning the piece of bubinga acquired on Sunday into some nice wood end caps. Oh, and get to work on the relay-based, microcontroller driven input switch... (This may turn into a bigger open-source hardware and software project.)

Quote:



Hmmm..........Constant current source plus buffering.
Edit: I do remember hearing the sound being articulate with the DBs underbiased ~44mV.

c0nsumer, Congrats! Max + Max = MAAAXX

A bit of the brown stuff is a great way to celebrate!

amphead, if you got access to a freq. generator and an oscilloscope you should compare input/output at different bias currents. I did this with a little with an application through my line in (page 63 on this thread).. not as good as proper equip, but nonetheless indicative... when I get access to equip, I'm certain to play around

c0ns: congratsX2! , and all the rest you do.. really cool stuff. good documentation too!

ruZZ.il, good point, I have access at my brother-in-laws workbench. He's an elevator controls manager, soon to be vice-president. Out of town at an elevator convention in Florida. When he gets back, I'll take a look on his workbench. You have brought to mind a good idea. You could take pictures of waveforms on your o-scope, with testpoint references and they could be posted on Tom's website. Edit: I have found that if I give my ears a 2 day rest and then re-evaluate the sound, component changes have a more pronounced effect. Kind of like a coffee bean taster who rinses with water between sessions! I think it reduces the noise floor and creates a black ambient space between notes within our eardrums/aural nerves.

when I have access to proper equipment, I may. meanwhile, my alien dac as a frequency generator and line in as an o.scope don't quiet cut it for me I guess the dac is decent as a source cause I'm more interested in comparing in to out, but my line in doesn't quiet cut it. it would be interesting to look at a whole bunch of responses though.. switching, non-linearities, etc... problem is, by the time I have access to the stuff.. I don't have the time to play. hmpf. I've actually had my eyes open for an old scope and f.generator for a while now.. btw, elevators are cool.. I like watching the relays switch stuff then manually push the things around sometimes muahahah pwnd. have fun with the equip! see if he has some export tools.. screenshot stuff or smtn, maybe you can update us

Do all international packages cost loads to ship from the USA as radiodaze have just quoted $23 for shipping to the UK. With parts for mine coming from parts express, michael percy, and my main mouser order il be paying close to $100 for shipping

Any idea if it's possible to get 2 pairs of 12fk6 tubes and x2 7-pin mini ceramic sockets with mount sent cheaper than $23?

Quote:



Try http://www.thetubestore.com/
I was quoted $7 for 3 tube pairs and 6 sockets with Air Parcel Post. I went with UPS though for $27.

So, would any of you have an interest in an audio input switch which can fit inside a Hammond enclosure along with the MAX?

I'm starting to work something up for my own use, and as soon as the microcontroller sample arrives I'll finish up the software.

This will be a full-on BSD-licensed software / creative commons board designs project, and I might sell PCBs and pre-programmed microcontrollers...

Quote:



For C4 and C5, I used 50V, 470uF caps.

You are right about the 12AE6/A. I got a chance to check with mine; the pair that is changing with the volume is 12AE6 non /A.

So that begs the question, what is different between the /A's and non /A's? or Is something wrong with my non /A tubes?

Forgot to add,

The range that I measured goes from 13.5V to ~18V (full volume).

Quote:



Cheers, but their 12fk6's are $7.95 each

Ive sent a message to another tube store, totally forgotten their name but they have a red background on their website and only sell tubes (no accessories).

Quote:



that would be cool! specially if its applicable to a lot of amps. I'd think it's even possible on a simpler system of gates, momentary switch, a 5v supply and some good relays.. a led or 2.. hmm but I'd really like to see something come along!

Quote:



My current thought is the following:

Two to five inputs. This is selected by soldering / omitting two jumpers.

Support for two buttons, one for next input, one for previous input. (You could just fit the next button if you didn't want to go back.)

Support for up to five LEDs, indicating which channel is active.

Microcontroller stores last input, so if you power the box off and back on it'll preserve the last known input. uC also allows for 200ms (I think) delay between breaking the old connection and making the new one.

4-pole relays so the switch can be used to switch balanced inputs. Pads for connecting things will support (but not require) screw terminals. Traces for this switching will be isolated from digital logic.

Simple on-board rectifier and power supply so it can be run from 5VAC - ~30VAC / 7VDC - 35VDC (or so). Then it can easily be fitted into many other designs. Or just run it direct from 5VDC.

Header on board for connecting switches and LEDs. This way the switch can be fitted in the front of a chassis, with the switch itself at the rear near the inputs.

All parts are common, through-hole, easy to find stuff.

Click to expand...

I'm thinking a single-piece rectifier, LM7805, small cap for smoothing power, PIC16F630, 5 (or 10) cheap NPN transistors, 8-pin header, five 4-pole relays, maybe some screw terminals, five resistors, bit of cable, one or two switches, five LEDs.

Uhm... that's about it, I think.

The most expensive thing here will be the relays.

Quote:



I don't think it's the tubes, per se, except that some tubes may have a propensity toward a stronger bass boom and perhaps more current.

Here's what I think is happening, but it's purely conjecture on my part:
I've been suspecting for awhile that the 470uf's may be marginal in the decoupling positions. CA4 and CA5 essentially ensure that the + rail and ground remain apart. In other words, they ensure that the ground stays at a lower potential than the rest of the circuit and also ensure that the + rail voltage doesn't dip from the power supply, through the CCS/tube, and on to the DB's. However, with the very powerful DB's in the output stage, it may be possible to overdrive the CCS through the tube. IOW, the high gain is sucking the current out of the CCS's, which are allowing the tube bias to climb as you turn the volume up.

Anyway, like I said - conjecture - we'll see if someone blows it full of holes. In the meantime, you might do all us a favor if you are able to change out those caps to larger values and see if those old 12AE6's do the same thing after that.