8 Channel NYD preamp Build Thread/General Questions

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

C12VR

Well-known member
Joined
Mar 13, 2021
Messages
242
Location
USA
Hello,
I am planning out an 8 channel build of the NYD 12av7 one-bottle preamp, and I am a little stuck on the heater supply. It seems, from my research, that a DC supply is best for a first-stage preamp tube. Having explored the possibilities for a linear supply and losing my mind to PSUD2 in the process, I renewed my search efforts and found that some forum members (Mr. Thompson-Bell, for example) use SMPS supplies for high current DC heater duties. 8 12av7 tubes in parallel would need 6.3v at .45A per tube, making for 3.6A total. The most optimal supply then must be around 10A and adjustable to 6.3v. Do I need any additional circuitry to adapt an SMPS to this use, e.g., LC filter? What should I know about SMPS power supplies thus implemented?
Edit: It seems the closest voltage is 7.5v; I guess a dropping resistor might work?
Is this supply OK? https://www.mouser.com/ProductDetail/MEAN-WELL/RSP-75-7.5?qs=JK6Bpmia/muWwvY9i8JHNw==
 
Last edited:
You might like to think about wiring the heaters for 12.6V rather than 6.3V. This means you can use a nominal 12VDC SMPS which gives you many more to choose from and it halves the current requirement to 1.8 amps. The one you linked to has a regular current limited output. These can often have trouble starting cold tubes because the short circuit current can be insufficient to warm the heaters enough for the resistance to increase enough for the supply to come out of current limit. Instead I always choose types that have a hiccup mode short circuit protection. This pumps the full rated current into the load for a short period, waits for another short period then tries again. I find these bursts of current sufficient to warm the heaters enough to exit short circuit protection mode if the supply is rated at twice the nominal heater current. So for your purposes you need a 12V supply rated at 3.6amps which is just over 43 watts. My favourites are the Meanwell LRS series. These have hiccup mode short circuit protection and I have found them to be exceptionally quiet. They do not need any external filtering for heater use. Something like the MeanWell LRS 50-12 would be ideal.

Cheers

Ian
 
You might like to think about wiring the heaters for 12.6V rather than 6.3V. This means you can use a nominal 12VDC SMPS which gives you many more to choose from and it halves the current requirement to 1.8 amps. The one you linked to has a regular current limited output. These can often have trouble starting cold tubes because the short circuit current can be insufficient to warm the heaters enough for the resistance to increase enough for the supply to come out of current limit. Instead I always choose types that have a hiccup mode short circuit protection. This pumps the full rated current into the load for a short period, waits for another short period then tries again. I find these bursts of current sufficient to warm the heaters enough to exit short circuit protection mode if the supply is rated at twice the nominal heater current. So for your purposes you need a 12V supply rated at 3.6amps which is just over 43 watts. My favourites are the Meanwell LRS series. These have hiccup mode short circuit protection and I have found them to be exceptionally quiet. They do not need any external filtering for heater use. Something like the MeanWell LRS 50-12 would be ideal.

Cheers

Ian
Thank you. At this point I must owe you at least a case of beer lol. In order to wire for 12.6v I would need to do a series string, right? Is there any disadvantage to this over parallel? Also, does this mean I can feed the output of the SMPS directly to the H+ and H- lines?
 
Last edited:
Thank you. At this point I must owe you at least a case of beer lol. In order to wire for 12.6v I would need to do a series string, right? Is there any disadvantage to this over parallel? Also, does this mean I can feed the output of the SMPS directly to the B+ and B- lines?
You do not need to do a series string. The 12AV7 can be wired for either 6.3 or 12.6 heaters. Check out its datasheet:

http://www.r-type.org/pdfs/12av7.pdf

Cheers

ian
 
Yes, all tubes with 12 in the American designation such as 12AX7, 12AV7, 12SN7 and so on can be operated at 12.6V. From my point of view this is even advantageous.

These tubes have two-part heaters, the series connection is made in the tube. Multiple tubes are then wired in parallel normally.
 
Last edited:
Yes, all tubes with 12 in the American designation such as 12AX7, 12AV7, 12SN7 and so on can be operated at 12.6V. From my point of view this is even advantageous.

These tubes have two-part heaters, the series connection is made in the tube. Multiple tubes are then wired in parallel normally.
Thanks! Ok, so basically the tubes are still being run in parallel, but the two heaters within the tubes are in series with each other. Great!
 
Screenshot (1).png
The B+ supply is standard linear design, very simple permutation of the original NYD schematic with a choke instead of the resistor. I simulated the load as resistive, using Ohm's law to derive an approximate cumulative resistance of 3750 Ohms (R=V/I=300v/.080A=3750 Ohm). Constant current of 80ma yields essentially the same results. Current estimation:
First stage: (235VDC -55VDC)/220k = 0.8 mA
Second stage: (235VDC-106VDC)/22k =6.1 mA
about 7ma, rounded to 10ma for security's sake.
Does this simulation look good?
Is there anything I'm missing or messing up?
I am also using a Hammond transformer with a 250VAC secondary at 130ma; is this acceptable?
The preamp schematic calls for 300v B+ rail, but I am expecting a slightly lower actual B+.
 
Last edited:
The schematic calls for 1/2w+ audio taper pots for the NFB/gain control, but since there is a cap blocking DC and it will be passing low-voltage AC signal, can I use a 1/4W potentiometer? Or is linear taper more or less fine for this app? I bought some nice Bourns pots with the right value and 1w rating, but I unwittingly selected linear taper ones, and Mouser doesn't seem to stock audio taper pots in conductive plastic or better beyond 1/4w. Also, can anyone weigh in on my previous question?
 
Also, can anyone weigh in on my previous question?
Hey C12VR, if you have questions regarding a circuit, please post them as well. It makes it easier to help you.

NYDAVE one bottle mic pre original.gif
The schematic calls for 1/2w+ audio taper pots for the NFB/gain control
Only half correct, the wattage refers to the static resistors not the potentiometer. You can use a 1/4 Watt pot, but audio tap is probably better than linear, just check it out what you like better.

My tip, before you build 8 channels of it, build one first to check the design. Personally, I'm not the biggest fan of the 12AV7 tube, I think it doesn't sound very good in this circuit (with negative feedback).

This variant of NYDave´s "OneBottle Pre" I like much more.
one bottle 6sn7.png
The B+ supply is standard linear design, very simple permutation of the original NYD schematic with a choke instead of the resistor. I simulated the load as resistive, using Ohm's law to derive an approximate cumulative resistance of 3750 Ohms (R=V/I=300v/.080A=3750 Ohm). Constant current of 80ma yields essentially the same results. Current estimation:
First stage: (235VDC -55VDC)/220k = 0.8 mA
Second stage: (235VDC-106VDC)/22k =6.1 mA
about 7ma, rounded to 10ma for security's sake.
You can make the calculation easier for yourself. Via R13 65 volts at 10k drop out, which means 6.5mA current consumption for the whole circuit. Your 10mA per channel estimate is on target though. (with a 12AV7, 6SN7 needs a little more)

Is there anything I'm missing or messing up?
I am also using a Hammond transformer with a 250VAC secondary at 130ma; is this acceptable?
The preamp schematic calls for 300v B+ rail, but I am expecting a slightly lower actual B+.
The 1H choke in your simulation is not that common, most I know (Hammond) are more between 4-10H for that expected current requirement. Which leads to an even better filtering. The resistive part of the choke will be above 50 Ohm.
Your transformer is a good match for 8 channels. 250VACxSQR of 2= 350VDC raw before ripple filtering. With just the choke your B+ will be over the needed 300VDC, I guess something around 320-330V. Thats cool, you can add another RC Filter for lower ripple or adjust R13. 20 Volts more or less are no problem in tube world, check you max voltage ratings of your capacitors, minimum 400VDC. The transformer and the 8 channels in one case, close to each other can result in direct hum problems. Heads up...
 
Last edited:
Awesome. I'll have to sell off the unfortunately expensive linear pots haha. Unless you have a use for 100k linear. I'd be glad to send some your way.
I just found some good audio taper at 1/4w which has the same rotational life as the Bourns at 1/4 the cost.
I've built a single channel of it, and it sounds good to me. Yeah, it was dumb of me not to post the schematic. The choke is a Triad c-24x. 1h with 1500V hipot and 230ma+ DC rated for 50ohms resistance. I thought of the hum, and figured resultantly that it was best to put the PSU in an in-line case. Unfortunately this raises the cost $100+ once the price of the box, connectors, and cable are put into play. Worth it for noise, I suppose.
Fun little experiment planned; 4 different makes of IPT: triple-shielded Newcomb 1:13, vintage Jensen Je-DB-E in reverse, Cinemag cmmi10C, Bogen 1:15, Lundahl ll1951 tiny high ratio like that Haufe you posted a few days ago. Once Igor at UTM makes those 1:20 transformers available again I might give them a try.
calculating the secondary loads will be a pain though.
 
Last edited:
Given that the B+ and H+ lines will consume almost 2 amps, is it safe to rate the mains fuse at 2.5A slow blow?
Also, should I put a fuse on the secondary of the transformer, or is this overkill? I see some people recommending this.
 
Last edited:
The choke is a Triad c-24x. 1h with 1500V hipot and 230ma+ DC rated for 50ohms resistance.
I hope the 1H is enough filtration, you may need an additional RC filter. You will see it.
I thought of the hum, and figured resultantly that it was best to put the PSU in an in-line case. Unfortunately this raises the cost $100+ once the price of the box, connectors, and cable are put into play. Worth it for noise, I
The challenge is the proximity of the power transformer and the signal transformers. Under certain circumstances, some channels will show too much mains hum due to magnetic crosstalk.
I might give them a try.
calculating the secondary loads will be a pain though.
Why? You define your desired input impedance and calculate the corresponding terminating resistor. This part is easy. If you need Zobel elements it can be a bit more complex, because they are transformer specific.

Given that the B+ and H+ lines will consume almost 2 amps, is it safe to rate the mains fuse at 2.5A slow blow?
Also, should I put a fuse on the secondary of the transformer, or is this overkill? I see some people recommending this.
I would definitely protect the secondaries of the transformer with fuses. Since you are planning to use an SMPS, you need to see what fuses it already has. I am not familiar with this. If you are heating with a conventional PSU, the inrush current may be a challenge for your H+ fuse. Maybe use a temperature dependent resistor or similiar.

Take into account the transmission factor of the mains transformer when dimensioning the primary fuse. The fuse that just survives the switch-on moment and then remains stable is the right one.😎
 
darn. Ok for safety's sake I will add an RC filter, and get a fuse for the transformer secondary. Jeez the expenditures never end do they.

So the secondary loads are just to define input impedance? How does the intrinsic impedance of the tube grid figure in?

yes, zobels are what I was really thinking of. I'll just try them no zobel except for the lundahl, which has a datasheet-indicated value for the capacitor and resistor.
 
Last edited:
The PSUD2 simulation claims that the 8 plates will draw 400MA at the transformer secondary. Is this an error, or do I really need a transformer with current capability that high?
 
The PSUD2 simulation claims that the 8 plates will draw 400MA at the transformer secondary. Is this an error, or do I really need a transformer with current capability that high?
I do not know all the numbers, but 400mA seems way to high for me. 8 amps (NYDave one bottle) with max. 10mA current demand is the real basis for calculations. The transformer you mentioned before is a good fit for your project AFAIK.
 
So the secondary loads are just to define input impedance? How does the intrinsic impedance of the tube grid figure in?
RG(rid) essentially determines the input impedance because the total impedance of a parallel circuit is less than the smaller AC resistance of the two parallel impedances. RTUBE is the intrinsic impedance of the tube grid.
20221211_133503.jpg
 
Last edited:
I do not know all the numbers, but 400mA seems way to high for me. 8 amps (NYDave one bottle) with max. 10mA current demand is the real basis for calculations. The transformer you mentioned before is a good fit for your project AFAIK.
Thank jesus and satan alike
 

Latest posts

Back
Top