Bijou All Tube Futterman Headphone Amplifier
Aug 3, 2009 at 11:08 PM Post #2,821 of 4,278
Yes, either the choke mod or the cap multipliers will maintain the voltage. It appears that some, but not all of the tubes pull a bit too much current, more than expected.

If I were designing the PS today I would use the cap multipliers and then no one would have regulation issues. I avoided them initially so that the amp would be as much "tube" as possible, including the older style RC filters. The only concession was the fet regulator. However, given that the RC filter is behind the regulator, it may as well be more sand.
smily_headphones1.gif
 
Aug 4, 2009 at 5:00 AM Post #2,822 of 4,278
the cap multiplier mod is cheap and easily done on veroboard - although i wouldn't recommend air-wiring it (attached photo is my first ill-fated attempt at air-wiring transistors in my original balanced Bijou for everyone to have a chuckle at
regular_smile .gif
).

if you have the space, it's easier to mount the cap multiplier off the PS board. i mounted it on the PS board, but it wasn't terribly difficult.

it's been awhile since i've heard my Bijou balanced, but Regal's mod definately improves on the dynamic abilities of the SE amp. much larger and bolder imagery too.

if anyone who wants to up the idle current of the input stage and add Regal's mod to an existing Bijou, do yourself a favor and buy a solder-sucker if you don't already have one. i assumed that fitting the larger 1W and 2W leads into the pcb pads might get tricky, so i ordered one of these from my local friends at Digikey: Digi-Key - K500-ND (OK Industries/Div of Jonard Ind Corp - DP-140). i'm glad i did. getting those thicker leads in was tricky. the best i could do was get one lead end all the wat through the pad hole, and the other, only half-way through the hole. i used Yageo resistors: Digi-Key - 30KW-2-ND (Yageo - RSF200JB-30K). i had a difficult time sorting through the lead diameters of various manufacturers' datasheets. YMMV.
 
Aug 4, 2009 at 5:24 AM Post #2,823 of 4,278
Aug 4, 2009 at 9:36 AM Post #2,824 of 4,278
Quote:

Originally Posted by regal /img/forum/go_quote.gif
I switched to the russian tubes with no further mods other than what was listed for the EC99's, the PS can handle it and the ruskies sound bettter.


Quote:

Originally Posted by holland /img/forum/go_quote.gif
Cap multiplier drops less voltage, as does the choke mods (I believe).


Quote:

Originally Posted by runeight /img/forum/go_quote.gif
Yes, either the choke mod or the cap multipliers will maintain the voltage.
If I were designing the PS today I would use the cap multipliers and then no one would have regulation issues.



Quote:

Originally Posted by fishski13 /img/forum/go_quote.gif
the cap multiplier mod is cheap and easily done on veroboard
if you have the space, it's easier to mount the cap multiplier off the PS board. i mounted it on the PS board, but it wasn't terribly difficult.



Quote:

Originally Posted by fishski13 /img/forum/go_quote.gif
also, one should increase the heatsinking on Q2 of the PS with the cap multiplier.



Wow
biggrin.gif
, thanks to all who replied, much appreciated.

Finally started my Bijou, soldered the P/S board last night, after the CTH I was surprised by just how few parts there are and how much room on the board there is. Will go with stock build for now, then when finished and I know everything is working will see if the P/S needs modding since results seems to vary.

The cap multiplier seems the way to go, but I don't think I am going to have enough room or the skills to make one on veroboard. Will have to see what parts are available locally first and then take it from there.

If I go with a choke supply which variation on the website is preferred?

And how big do I need to go on the Q2 heatsink, the one I have is 2" tall and I can't go taller, will that be enough?
 
Aug 4, 2009 at 2:21 PM Post #2,825 of 4,278
Both choke supplies will do the same think, but variation 2 is a bit easier to do. You just eliminate R7, R8, and C5 and insert the choke leads into the right pads for R7 and R8.

A 2" heatsink should be enough. The fet will get hot though.
 
Aug 4, 2009 at 5:09 PM Post #2,827 of 4,278
Heh, I have the same problem with space, but really running @ 245V is no big deal for me.

I thought of 2 approaches. Changing the CRCRC filter to reduce the drop (reduce R, increase C). Using 220ohm instead of 330ohm resistors and increasing the big Cs to 680uF, for example. I think ripple is a bit worse off, not sure, haven't really looked at it beyond that.

Using a semi-conductor bridge instead of a tube rectifier will definitely work. I think you can use one of the copper caps in place of the EZ81 and have it custom made by that company. I can't recall who though. It's simple internally, just using R's to emulate the voltage drop and create a soft-start situation (not as soft as the EZ81). If you make your own, you can use a thermistor.
 
Aug 5, 2009 at 3:52 AM Post #2,828 of 4,278
holland - If I can get 245V then I will probably go with that as well. My space problems are vertical, I have room horizontally for the cap multiplier but on the wrong side of the P/S board, will play around with the layout I have planned and see if I can swap the board around.

Since you are using the same Rcore transformer what is the best way to use the two sets of 6.3V secondaries? I had planned to use one for the P/S board and the other for the amp boards to keep things simple, is there a better way?
 
Aug 5, 2009 at 4:40 AM Post #2,829 of 4,278
Forte, I put them in parallel and added a pair of series resistors to drop the AC voltage down some. Paralleling is easy, connect the center taps, that will create a reference point in the middle and you can probe the wires with a DMM to see which wires to connect.

One thing to note, unless you do Amphead's heater mod, is that the heater is biased at some relative DC level. So parallel would be best for the stock configuration without Amphead's heater mod. Though I think if you keep the center taps connected together you can wire them separately. I haven't thought much about that. It was easiest for me to use a single heater configuration as it required less resistors to tweak the voltage.
 
Aug 5, 2009 at 5:52 AM Post #2,830 of 4,278
Quote:

Originally Posted by Forte /img/forum/go_quote.gif
holland - If I can get 245V then I will probably go with that as well. My space problems are vertical, I have room horizontally for the cap multiplier but on the wrong side of the P/S board, will play around with the layout I have planned and see if I can swap the board around.

Since you are using the same Rcore transformer what is the best way to use the two sets of 6.3V secondaries? I had planned to use one for the P/S board and the other for the amp boards to keep things simple, is there a better way?



i wouldn't worry about the distance of wiring the cap multiplier to the PS board. others smarter than me please chime in, but i don't think any relatively small extra lengths in the PS circuit would be an issue with respect to noise or resistive DC output.
 
Aug 5, 2009 at 5:56 AM Post #2,831 of 4,278
anybody,
how do i calculate the output impedance of the Bijou into a known load, both on paper and in situ DVM readings? i have no NFB.
 
Aug 5, 2009 at 6:10 AM Post #2,832 of 4,278
Aug 5, 2009 at 6:14 AM Post #2,833 of 4,278
Quote:

Originally Posted by fishski13 /img/forum/go_quote.gif
i wouldn't worry about the distance of wiring the cap multiplier to the PS board. others smarter than me please chime in, but i don't think any relatively small extra lengths in the PS circuit would be an issue with respect to noise or resistive DC output.



Yes, a little bit of extra wire won't matter here as long as it's not too long.

Fishski the Zo of the Bijou is about 55R with no NFB.
 
Aug 5, 2009 at 6:21 AM Post #2,834 of 4,278
Quote:

Originally Posted by runeight /img/forum/go_quote.gif
Yes, a little bit of extra wire won't matter here as long as it's not too long.

Fishski the Zo of the Bijou is about 55R with no NFB.




thanks!!!, but how do I go about getting 55R?
 
Aug 5, 2009 at 7:36 AM Post #2,835 of 4,278
Ah. I did this calculation quite a while ago and I don't remember how to do it again, but the output impedance is given by:

Zo = rp / (2 * (1 + mu))

The ECC99 has an rp of 2300R and a mu of 22 so ...

Zo = 2300 / (2 * 23) = 50R

The 6N6P has an rp of 1800R and also a mu of 22 so ...

Zo = 1800 / (2 * 23) = 39R

In this configuration, with the crossed outputs from the phase splitter, both O/P triodes operate as cathode followers. So the Zo is essentially the paralleled Zo of two cathode followers.
 

Users who are viewing this thread

Back
Top