Tube DI Box

[Tubetec]

I did see a Max 1771 board with just HT on ebay , its the size of a postage stamp , costs around 4 euros ,
If you need an elevated heater supply , it seems a seperate winding on the transformer is needed ,that HT/LT board is no good .

 

[Wavelength]

Ian,
2 things, the inrush current on a filament can sometimes exceed the max current and shut down the DCDC portion of the converter. Another thing to watch as I got burned with a 12.6V supply and dual 6GM8/ECC86 in series is the filaments are never the same so one will draw more than the other. So you may want to look at using a doubler chip as they have a method of creating a mid point and that is what I used to make sure each had their own 6.3V supply.

 

[ruffrecords]

I did see a Max 1771 board with just HT on ebay , its the size of a postage stamp , costs around 4 euros ,
If you need an elevated heater supply , it seems a seperate winding on the transformer is needed ,that HT/LT board is no good .

Do you have a link to it?

Cheers

Ian

 

[ruffrecords]

Ian,
2 things, the inrush current on a filament can sometimes exceed the max current and shut down the DCDC portion of the converter.

I use SMPS for 12V heaters all the time. I always rate them for twice the nominal heater current and pick ones that have hiccup mode. This way I find I can guarantee they will fire up into cold heaters. In this particular case we have two SMPS in series; the AC/DC one in the wall wart and the DC/DC that produces HT and 6.3V. The 6.3V output is rated at 1.5 amps but I found it will only fire up into a single cold 300mA heater (ECC88). I suspect the problem is the wall wart is rated at 1 amp and is being asked for current pulses much bigger than that. What I will do is monitor the wall wart output to see if it goes into current limit with cold heaters.

Cheers

Ian

 

[nsalis]

Based on the Philips data sheet about the ECC88, this tube can only be used with a heating voltage of 6.3 volts (between pin 4 and pin 5). Pin 9 is the shield between the two triodes and not the center tap of the two heaters connected in series. It is therefore not possible to operate the heating filaments of the two triodes separately.

The heater current is 365 mA (not 300 mA).

The eBay device you mentioned can supposedly deliver 15 mA at 280 VDC and 1500 mA at 6.3 VDC.

The total output power is therefore:

PPlate voltage = U*I = 280*0.015 = 4.2 W.

PFilament voltage = U*I = 6.3*1.5 = 9.45 W.

Let's assume that the entire device causes a heat loss of 20%.

The supplying power supply must therefore provide:

(4.2 W + 9.45 W =) 13.65 W. This corresponds to (100 - 20) = 80% of the power consumption. 100% is therefore ((100*13.65)/80 =) 17 W.

The supplying power supply must therefore provide a current of I = P/E = 17/12 = 1.4 A at an output voltage of 12 VDC. As others mentioned, the heater starting current is more than the nominal value; I would at least double the output power of the power supply, i.e. specify a power supply with an output of 12 VDC / 2.8A.

Your 12 VDC / 1A power supply is definitely too weak and needs to be replaced with one with around 2.8A or more.

e.g. Meanwell GSM36B12-P1J

https://uk.farnell.com/mean-well/gs...ical-ac-dc-12v-3a/dp/2815798?ost=gsm36b12-p1j

or

https://tinyurl.com/5av8sh6j

Nick Salis

 

[ruffrecords]

@Nick many thanks for the analysis. It is much appreciated.

The ECC88 just happens to be the name of the spice model I use. In practice I usually use 6922 tubes which do have a heater current of 300mA.

Although the eBay device is rated at 1.5 amps it will not reach anywhere near that amount in this project so a more modest mains power supply will be adequate.

Nice choice of MeanWell power supply. I generally use their enclosed types but for this project, the type you selected is ideal. Thanks

Cheers


Ian

 

[Tubetec]

that one ,
https://www.ebay.ie/itm/404197658522

Another option here , all through hole components ,
https://www.ebay.ie/itm/174215780131

 

[ruffrecords]

that one ,
https://www.ebay.ie/itm/404197658522

Another option here , all through hole components ,
https://www.ebay.ie/itm/174215780131

Thanks for the links. I ordered one of each. I like the through hole one because it seems to have an extra filter stage (as shown in the chip datasheet) which drops the ripple a lot but it will still need some more RC smoothing after that to get the ripple down to levels acceptable for audio use.

Cheers

Ian

 

[Tubetec]

Yeah I noticed the extra filter on the through hole board ,
Another bonus is the through hole PCB can output upto 300v ,
only downside is it doesnt have the ten turn pot which gives nice precise adjustment of your HT .

 

[ruffrecords]

All the circuits I have evaluated so far have used the ECC88/6922 tube because it has a low enough rp to be able to drive a current into a 2K4 load (600 ohms reflected via 2:1 transformer)

However, the ECC88 and friends are all 6V heater only types of tubes. If we use a 12V supply, things would be simplified if we used a tube with a 12V heater. So I decided to do a few sims using my least favourite tube (because of its high distortion) the 12AU7 running from a 200V supply (to allow sufficient voltage drop to reduce HT ripple to suitably low levels using a simple RCRCRC filter). The designs were all adjusted for a nominal quiescent current close to 5mA.

I tried several topologies but the best two were the SRPP and the mu follower both of which could drive a reasonable level into 600 ohms. The SRPP spectrum was almost entirely second harmonic but at the test level the distortion was quite high (~5%). The mu follower distortion under the same conditions was significantly lower (~0.3%)but did contain a harmonics from the 3rd upwards. so the two designs will undoubtedly sound different. I tried two different 12AU7 models and found there was a 2:1 difference in distortion between them und er identical conditions which just shows how little you can trust triode models whn it comes to distortion.

The mu follwer has the advantage that it does have about 14dB of gain compared to the 4dB of the SRPP.

It is quite easy to devise a simple PCB that can be used to build either design so I shall make a start on that soon.

Cheers

Ian

 

[Tubetec]

As I see it ,
the compound triode is only an attempt to mimic an impedence ,
How would an optimised inductor , as found in the Tele V series fare in the sim ?
and how does it relate subjectively to what we hear in the real world ,

 

[ruffrecords]

As I see it ,
the compound triode is only an attempt to mimic an impedence ,
How would an optimised inductor , as found in the Tele V series fare in the sim ?

I am not familiar with the Tele V. Do you have a schematic you could share?

and how does it relate subjectively to what we hear in the real world ,

Judging by the number of tube tone boxes I have built for people, whatever it does, the SRPP sound is well liked. The guy who purchased my very first tube mixer which uses 6CG7 mu followers throughout, uses it to process and mix three final stereo stems.

Cheers

Ian

 

[rock soderstrom]

I am not familiar with the Tele V. Do you have a schematic you could share?

I guess he means the Telefunken V76/V72?

 

[merlin]

The mu follwer has the advantage that it does have about 14dB of gain compared to the 4dB of the SRPP.

The mu follower is intrinsically a gain stage plus a cathode follower, so it would make more sense to use a gain stage plus cathode follower in cascade, since you are voltage limited not current limited in this case.

 

[ruffrecords]

The mu follower is intrinsically a gain stage plus a cathode follower, so it would make more sense to use a gain stage plus cathode follower in cascade, since you are voltage limited not current limited in this case.

Except you lose the inherently low distortion of the mu follower and replace it with the regular CC stage distortion plus you probably double the HT supply current requirement but you are right, it will have a better voltage swing.

Cheers

Ian

 

[merlin]

Except you lose the inherently low distortion of the mu follower

Not if you bootstrap the gain stage anode resistor -it's literally the same circuit laid out horizontally rather than vertically. Current doesn't need to double since you have the freedom to run the gain stage at low current. Anyway I thought tube character was the whole point of this DI?