just a "One Bottle" mod

neodyms said:

this the pdf

sorry

Click to expand...

Thanks for posting and sharing.

Agree on the previous suggestion increase the fb pot to 470K and Kingston's other notes on transformers.

I'm not 110% convinced about your tl783 phantom regulator.

You are only supplying 2 channels of phantom at say typically around 4-6mA each = total of around 10-12mA.

You currently have a worst case no load of 1K+2k7+82R which is around 12mA. If the 1K trimmer is at minimum you'd have 2k7+82r for 48V = 17mA.

With 2 mics connected you have a worst case high load of less than 50mA (17mA+2*48/6k6||6k8).

1. 4*220uF looks way too much for a load of only 50mA for 2 channels. Maybe this was a design to supply more channels? So is there too much capacitance on input?

You'd probably be better off using CR filters, rather than lots of parallel capacitors. 4 high uF caps will cost quite a bit relatively speaking.

2. The TL783 needs 15mA to maintain regulation (from the data sheet) so the output voltage may not be regulated to 48V under very low loads.

Add on one single low power mic like an AKG 414 which only needs 2mA and you only have 14mA-19mA total in the worst case. I know they suggest 82R for the adjustment resistor, but I'm not sure that's enough. It might just be enough to make the TL783 work. Then again it might not.

Suggested changes:

Change the 220uF +220uF + 220uF+ 220uF to a simple single 220uF + 120R + 220uF CRC filter, The maximum permissible value for Rfilter (120R) will depend on your transformer no load voltage (TL783 needs 10-20V across it for good ripple rejection).

Simply "throw away" a tiny bit more current (3mA more) by connecting a 15K 0.25W resistor to ground from the output of the regulator. Or perhaps reduce R44 to 68R and R46 to 2K2. Then you'll always have at least 15mA-20mA flowing even without a mic connected, so the TL783 will always regulate.

Some aspects of the phantom power are overkill. Four by 220uF smoothing is unnecessary for a single channel mic pre. A single 470uF will be plenty. Even a single 220uF would give little more than 150mV of ripple. Also, the original circuit uses a 40VAC secondary which gives at most  8V across the TL873, which would drop to 3V is the mains dropped 10%  - not really enough IMHO which is why I recommend a 50VAC secondary which will give at most 22V across it and over 15V even at -10% mains. You might also want to think about a heatsink for the TL783 just to be on the safe side.

Cheers

Ian

neodyms said:

You could even take the feedback loop from behind the output coupling cap and skip the separate feedback cap completely (saves one potentially expensive film cap).

Click to expand...

i am in a quality research more than economical compromises ; what sound better ?

Click to expand...

I'm almost certain they will sound the same. The separate feedback has marginal impact on low frequency impedance. I suspect it can just barely be seen on FFT display, but can't be heard.

Maybe this was a design to supply more channels?

Click to expand...

you right

------


about the DI input impedance matching with the microphone input ... no problem ?

ruffrecords said:

Some aspects of the phantom power are overkill. Four by 220uF smoothing is unnecessary for a single channel mic pre. A single 470uF will be plenty. Even a single 220uF would give little more than 150mV of ripple. Also, the original circuit uses a 40VAC secondary which gives at most  8V across the TL873, which would drop to 3V is the mains dropped 10%  - not really enough IMHO which is why I recommend a 50VAC secondary which will give at most 22V across it and over 15V even at -10% mains. You might also want to think about a heatsink for the TL783 just to be on the safe side.

Cheers

Ian

Click to expand...

Agree on the 50VAC transformer. That's often available as a bias tap on e.g. Hammond 300 series products. Looks like a good change from 40VAC. The tl783 can cope with high voltages, and needs quite a big voltage drop to work.

Worst case in this design (2 channels) would be 50VAC transformer -> 22V drop over TL783 @ 50mA into two short circuits (48V into 4 off 6K8 in parallel + Ino_load). That gives just over 1 watt dissipation. Thermal resistance junction to ambient is 25C/watt. And max operating junction temp is 150C. So I think it's pretty safe without a heat sink unless something goes very horribly wrong.

ok corrected !


again, what do you think about the DI input impedance matching with the rest of the microphone input ... does it work ?

neodyms said:

ok corrected !

again, what do you think about the DI input impedance matching with the rest of the microphone input ... does it work ?

Click to expand...

For the 48V you need EITHER the change of R44 to 68R OR add the 15K. Not both.

As for DI, I've seen various schemes.

The SE5a of ampmaker just connects the guitar directly to the grid with a 1M self-bias resistor, exactly as you have shown. http://www.ampmaker.com/images/ak01kit/ak01sc2.jpg I built that amp for my brother and it worked fine. Some people report noisier volume controls due to some grid current leakage.

Gyraf G9 uses a 220n coupling cap http://www.gyraf.dk/gy_pd/g9/g9_sch.gif I also built a G9 and used it with my bass OK. Keep the leads to the input jack short and screened to avoid oscillation.

The updated Mila1 with DI has a 270R grid stopper. http://musiciansgig.com/mila/MILA1-Updated.pdf YMMV.

thank a lot every body

any other suggestions ? do you it could be built like this ?

wich brand use for i/o transformers ?

neodyms said:

thank a lot every body

any other suggestions ? do you it could be built like this ?

wich brand use for i/o transformers ?

Click to expand...

Watch out! Your DI "improvement" will probably cause you major problems. The grid of the first tube must be connected to ground via the 1M resistor for DC in order to properly cathode bias. Check out the G9 schematic carefully again for the correct way to do this. I cannot guarantee the exact capacitor value either although it should be about right. If you happen have a bigger film or pio capacitor you could probably use that too. e.g. 1uF. I would avoid an electrolytic capacitor here.

People have had a lot of success with Edcor output transformers. wsm10K/600 should be about right according to what I've read and tried myself. Input transformers are much trickier. It is difficult to get hold of good 1:10 or 1:15 input transformers. There's always evilbay. But don't pay too much and make them too expensive for the rest of us.

If you make simple mistakes like this with the high voltage supply

Click to expand...

of course that's very dangerous, i am aware about high voltages, i often repair guitar amp and know the precautions to take and the risks 

----------

about this project i am just trying to be nearliest from the original NYDave one bottle schematic.
i think i will start this project by the PSU then the signal board and experiment values in function of the sound that produce.
i will send some reports with advancement of the project.

wich brand use for i/o transformers ?

Click to expand...

For input, you may already be aware of this - you need something shielded/potted.  Most grid transformers already are, but in case you were considering one of the Edcors for this job, know that that may be inviting hum pickup issues.  Sowter, Lundahl, OEP, Jensen, Cinemag, numerous vintage types - all good.  May want to comb the forum for specific user feedback to find what might best fit your tastes. 

Edcor's are a great choice for the output and it's hard to beat their price vs performance number.



Do recheck your DI arrangement.  I believe you want a simple 2 conductor shorting (Normally closed) jack there.  It also appears that you have the .22 cap in the wrong position - look at the G9 schem again.  You may not even want or need that cap ( C20 ).

For the PSU transformer , how many VA do i need ?

2x24 mA=50mA x 240v = 12 VA
2x600mA=1,2A x 6,3 v = 7,56VA
PLUS phantom PSU 
50VACx50mA = 2,5VA

+30% buffer = 22x1,33 = 30VA


Is it right ?

neodyms said:

For the PSU transformer , how many VA do i need ?

2x24 mA=50mA x 240v = 12 VA
2x600mA=1,2A x 6,3 v = 7,56VA
PLUS phantom PSU
50VACx50mA = 2,5VA

+30% buffer = 22x1,33 = 30VA


Is it right ?

Click to expand...

Normally you need to look at current per transformer winding. The overall wattage is rarely/never an issue for (high voltage) tube transformers unless you're building a power amplifier. Heater current on the other hand, often adds up very quickly. Phantom will only ever reach 50mA under fault conditions. More likely to be around 30mA. AFAIK the 12av7 will be biased at around 10mA + 1mA per channel [106V 1.8V cathode, 55V 2k2 cathode resistor]

So that would be:
winding 1a 240V rms (10mA + 1mA) *2 = ±22mA
winding 1b 50V bias tap for phantom (2*6mA + 15mA) = ±27mA [unless the phantom was short circuited permanently]
Total HT winding = ±50mA
winding 2 6,3V heaters @450mA *2 = 900mA

You might just get away with a Hammond 370AX, but I think that might be pushing it a little. A typical transformer choice that I would make would probably be something like a Hammond 370CAX, but if you wanted to play really safe then a 370dax would certainly do the job and some more for a couple of euros more. See the Tube Town store at http://www.tube-town.net/ttstore/index.php/cat/c155_300-Series-Classic-Tube.html I have no connection with them, but have had very good service from these guys in the past. Obviously you may want to shop around. Or you could look at splitting off the phantom and going for a 370AX and a separate phantom power transformer, or an external 48V smps wall wart, or indeed a voltage tripler, like on the G9. Depends how big your case is, and if you want to add some more channels. But I would generally prefer the bigger transformer rather than fancy electronics. By the time you pay for shipping and mounting up of 2 transformers, you're seldom better off. Bigger transformers also regulate better. If you want to use a choke in the power supply, a Hammond 193C could be a good choice.