Tube DI Box

[Tubetec]

In my new PSU I avoid the IEC socket ,the tooling and the extra space it occupies by using metal cable glands .
I usually just fit a 3 pin plug at the end , but you can also fit an IEC input if you prefer that .

One studio I worked in had all its rackmout gear powered from IEC plug boards ,
So it was mostly IEC male to female cables , it saves a lot of space compared to standard Irish/UK plug boards and sockets .

A Hammond diecast box is costly nowadays , its also not as good as steel for magnetic screening
Ive settled on a standard metal clad surface mounted dual gang back box ,
It costs around 5 euros and blanking plates are also available .
Im using two boxes bolted back to back ,
One compartment for HT the other LT .
Theres no need for a PCB , everything can be hardwired with a few tag strips mounted on brass pillars.

 

[merlin]

I am -- NOT -- "bragging" or saying "I'm great" or anything like that, which other members on here believe that is what I am doing!!! NO!!!.....
How many other members on this forum could design the "transportable equipment racks" ??? YOU???

 

[beatnik]

Another idea for the chassis could be something similar to the old DBX 160 VU

This unit could be used as desktop / standalone with wooden sides, or the sides could be removed and two units joined together into a 2U 19" system

 

[ruffrecords]

@ruffrecords ...

Ian....forgive me for even mentioning the 1 RU enclosure concept. I was merely commenting on what someone else mentioned in this thread....

No problem. No reason why a DIYer couldn't build it in a 1U box. I see this more as a project than a product - one that could be built in several different ways.

I need to look up the dimensions of a 12AX7..... etc.

In my mind, the gizmo box has to be maybe 3" tall to enclose the tube(s) mounted vertically. Fine....so be it. The box needs to be built from cast iron <lol!!!> to be able to withstand being stepped on the floor of the studio.

the 12??7 series of tubes are generally 2 3/16 inches tall.By the time you add a socket, PCB and clearances you are approaching 3 inches. They will definitely fit into 70mm

We all seem to agree that wall warts and line lumps SUCK <g> so the Magic Box needs to have an IEC power inlet. Here it gets a bit goosey for worldwide usage.....120/240 VAC?? 50/60 Hz??

Perhaps the most straightforward approach to that would be a custom transformer with international primaries. If it does not already exist it would be a useful component for lots of builds. (Carnhill VTT2302 ??)

Cheers

Ian

 

[PermO]

I Like this form factor for something active single tube DI thing;

https://www.reichelt.nl/nl/nl/deskt...tph-1850-150-p126234.html?&trstct=pol_8&nbc=1

Not the cheapest, but it does have a single sliding flat frontpanel that will be easy to customize.
Great as floorbox and desktop unit.

In doubt now as on second look it seems the front panel is fixed to the extrusion parts

 

[Wavelength]

All,
When I design something like this I like to start with either power or the output.
Power:
a) Phantom, problem low current low voltage, not great for anything tube related.
b) 12vac in and then a step up, you could do pretty much. Only negative is if the wall wart dies.
c) 9Vdc Boss negative center, I probably have 50 of them sitting around, common step up designs, good for anything with a 6.3V heater
d) 12Vdc same idea as above works with 1/2 the current as (c) but has the same problem as (b).
e) AC input pretty much do anything. I have a nice switcher in front of me for a pedal board amp: 6.3V@4A, 275V@140ma, also have a bunch of small power transformers from Hammond or Heyboer that would work really well: 370Vac-CT @ 30ma DC, [email protected] AC.
Output:
a) 10K:600 problem here is you have to have high transconductance and a bit of current with a low Rp to make it work. Most direct current units don't work as well as parallel feed which then has a few more parts to it.
b) 600:600 Cathode follower idea using either a current source or an BAC (big ass choke) from Cathode to ground then cap coupling the output transformer.

Enclosure comes after you make those decisions. Forgot I have the Rupert Neve DI thingy here and I think a couple others.

 

[PermO]

 

[RogerAF]

You might look into the Korg NuTube. I came across these a decade or so ago. It is based on a vacuum fluorescent display that has the characteristics of a triode. It runs on very low voltage and current. I had not seen one in action until recently when I saw a video of Jennifer Batten showing off her rig. Her amp was made by some tech in Germany and has a floor unit preamp that has many tones and uses the NuTube to drive a class-d power amp that she's very happy with--because it weighs hardly anything (especially when compared to a Marshall!) If you aren't familiar with Jennifer, she toured with Michael Jackson and then with Jeff Beck. She's a marvelous guitar player with a huge variety of tones and sounds. I thought the rig was quite something. I think you could definitely power this with a phantom supply.

 

[Tassieviking]

Just run it from a normal pedal supply, 9-12V is good enough to run the heaters in any 12A*7 tube.
If you want some inspiration for ideas then just look at Sushi Box pedals, he also sells PCB's for some DIY tube pedals.
His Nobelium pedal has a built in DI and all his pedals run on 9-12V and the tubes run on about 240V for the DIY ones.
https://www.sushiboxfx.com/
https://c2celectronics.com/

 

[ruffrecords]

I favour a submini tube mounted up in an XLR to jack adapter as cathode follower ,
Remote supply/ line out stage .

Id like to ask Ian to repeat the ECC88 CF test into 1M ,

I ran the schematic as posted with a 600 ohm load and a 1Meg load. The only real difference was the change in the low frequency peak due to the change in damping due to the load.. At 1V peak input, the distortion was 0.21% into 600 ohms and 0.17% into 1Meg. mostly second harmonic and a little 3rd as expected. At 10V peak input the distortion was 4.8% into 600 ohms and 3.25% into 1Meg. Lots of harmonics from the third upwards so it is close to clipping.

Note that the circuit is not a regular CF. We are not wanting to output +26dBu into a 600 ohm load. All we want to do is drive a 10K line input or a 600 ohm mic input. So we don't need to run the cathode at around half the supply volts to obtain a wide signal swing. As designed, the sim has the cathode sitting at 7.8V. We don't have 100% feedback, we have just 50% because the two resistors in the cathode are equal. This lets a little more of the tubes 'tone' through to the output. Even so, it takes a +20dBu input to get it close to clipping. Note also that we are not trying to drive a 600 ohm load direct from the CF but via 4:1 step down transformer. The load seen by the CF is therefore no more than 9K6.

Cheers

Ian

 

[Tubetec]

Thanks Ian ,

 

[ruffrecords]

Just run it from a normal pedal supply, 9-12V is good enough to run the heaters in any 12A*7 tube.
If you want some inspiration for ideas then just look at Sushi Box pedals, he also sells PCB's for some DIY tube pedals.
His Nobelium pedal has a built in DI and all his pedals run on 9-12V and the tubes run on about 240V for the DIY ones.
https://www.sushiboxfx.com/
https://c2celectronics.com/

Thanks for the link. Much appreciated. As I suspected he uses the" standard" NE555 based Nixie drive HT SMPS and a 7806 for heaters. This also confirms my view that 12VDC supplied tube stomp boxes are well established.

Cheers

Ian

 

[Wavelength]

Ian,
What schematic are you talking about? So if this is a CF into a 10K then we are talking about a 4:1 loss of signal.
Roger,
Like the 6418 the NuTube requires a lot of solid state to work. It's also really really microphonic. I know the designers of that tube and Pete Millet has done a ton of work with it. It seems all the big companies that went there have killed that product. Probably too many returns.

I need a DI May 1st for a show I am doing in Florida so I worked this up today in like an hour. I know solid state but the tube stuff was going to take too long. Still interested but I wanted a 9V option. BTW VCC = 8.5V, it's a shunt regulator I use on my FUZZ pedals.
When the stomp LED is off the output is muted and the input has a 1M stopping resistor so you can plug in and not worry about blowing the LSK189. It has a switch for phase, pin 1 grounding and input cap which should not be needed in 99% of the cases.
Boards are already in progress at JLPCB.

thanks,
Gordon

 

[ruffrecords]

Ian,
What schematic are you talking about? So if this is a CF into a 10K then we are talking about a 4:1 loss of signal.

The one in post #40.

Cheers

Ian

 

[MidnightArrakis]

A Hammond diecast box is costly nowadays

>> Then again.....the "diecast box" module within this enclosure design program is available to use as well:



And.....if you ever happen to use PELICAN cases and need to have a panel designed and placed within one, there's this:



/

 

[Wavelength]

Ian,

Ok ran TubeCAD on the schematic in #40. Gain = 0.9, F3 = 15.8Hz, Zout = 84.5ohms, PSRR = -31.5dB, with the 10K/600 gain goes to -13dB or 0.220, Vout max -7/+4.89.

You know you might be better off with a 600:600 transformer your gain becomes 0.8. You would need a larger cap 15uF gets you back down to 15.7Hz.

If you used a 6SN7 or the single ones like 6C5 or 6J5 or the 12V6 versions VT133 boost the B+ to 250V and stuck that 60H/8ma choke from cathode to ground into a 15uF cap then a 600:600 you would have pretty much a 1:1. Vout max would go to 15.556Vac rms.

Thanks,
Gordon

 

[ruffrecords]

Ian,

Ok ran TubeCAD on the schematic in #40. Gain = 0.9, F3 = 15.8Hz, Zout = 84.5ohms, PSRR = -31.5dB, with the 10K/600 gain goes to -13dB or 0.220, Vout max -7/+4.89.

That seems close to what to get in LTspice

You know you might be better off with a 600:600 transformer your gain becomes 0.8. You would need a larger cap 15uF gets you back down to 15.7Hz.

Define better off?

If you used a 6SN7 or the single ones like 6C5 or 6J5 or the 12V6 versions VT133 boost the B+ to 250V and stuck that 60H/8ma choke from cathode to ground into a 15uF cap then a 600:600 you would have pretty much a 1:1. Vout max would go to 15.556Vac rms.

Thanks,
Gordon

Except we do not need anything like 15V rms output capability.

Cheers

Ian

 

[Mondy]

Would like to thank Ian for looking into this, Tube DI's have always been a fascination for me and I am still on the lookout for a good one for live use.


I have always been intrigued by this “"Williamsonic"” design ...

http://www.williamsonic.com/TubeDI/

Seems to be very intelligently designed, it being used for Double Bass duties and the fact that the designer does not over hype the design and performance of the device, gives me confidence in its capabilities. I have not had a chance to build and try though. I don’t think I would be bothered with the triaxial input setup, just regular cable input.


Could a circuit based on this topology be worth considering?

Cheers
Ray

 

[ruffrecords]

Would like to thank Ian for looking into this, Tube DI's have always been a fascination for me and I am still on the lookout for a good one for live use.


I have always been intrigued by this “"Williamsonic"” design ...

http://www.williamsonic.com/TubeDI/

Seems to be very intelligently designed, it being used for Double Bass duties and the fact that the designer does not over hype the design and performance of the device, gives me confidence in its capabilities. I have not had a chance to build and try though. I don’t think I would be bothered with the triaxial input setup, just regular cable input.


Could a circuit based on this topology be worth considering?

Cheers
Ray

That is an intriguing design. At first I could not work out what he meant by a cascode follower until I realised it is just a regular White follower. This is clearer if you look at the Jensen app note the design is drawn from. I also note that Jensen specify a 10K:600 output transformer which is just what I was planning to use. Interestingly he uses a Lundahl mic input transformer backwards for the output transformer -something I have also seen used elsewhere. The strange thing is the White follower provides a very low output impedance and good drive capability which he then throws away by feeding the output transformer via 470nF and a 1K series resistor.

However, I have been wondering what to do with the other half of the ECC88 I was planning to use so using it to make a White follower is definitely of interest. I will do some sims and report back. Many thanks for the link.

Cheers

Ian

 

[MidnightArrakis]

>> The "Williamsonic" circuit requires a -- triaxial BNC-type -- of input connector and that baby costs $30!!! OUCH!!!

/