Tube-saver (non-redplating) circuit

[R.G.]

I was thinking some more about the power tube monitoring circuit I mentioned in the thread about the polyfuse for stopping redplating disasters. It connected up with some prelim work I'd already done.

It looks like this is possible:
- 10R 1% cathode resistors to sense cathode current
- Tube current read by voltage across the resistors
- One bias pot per tube
- One RGB LED per tube indicating below (blue), at (green), or above (red) target current
- Ability to open B+ to the tubes on overcurrent sense
- Pot to adjust target current to match the intended tube (i.e. EL84/6L6/EL34, etc.) and operating point
- pot to adjust target overcurrent disaster/shutdown current
- pot to adjust the time an overcurrent can exist before shutting down
- LED indicator of which tube had an overcurrent
- Reset button for resetting the thing back to active after an overcurrent
- Bypass switch for turning B+ back on in case the guitarist decides he's smarter than the circuit

That function can all go in one $2.00 chip. The pots, LEDs and a box to put it in cost a lot more.

In fact, the packaging is the real challenge. The electronics is pretty simple compared to how to package this so it could be added to an amp and be both non-intrusive and useful.

[martin manning]

I think something like this has been done already in a product I saw a couple of years ago. Maybe this one, but I'm not sure. http://www.tentlabs.com/Components/Tube ... age24.html

Another: http://www.mystere-usa.com/index.php?op ... Itemid=174

[R.G.]

I'm a little shocked that they don't exist in quantity.

From what I read, those things do a little different job. They actually do all the biasing themselves. I developed a circuit that did that for the Workhorse amps, but when I looked at what it really did, the Red/Green/Blue stoplight indicator system was so much simpler that we ditched the autobias for the stoplight system. The problem was that tubes do drift around as they heat up and age, so the circuit has to be tinkering with the bias all the time, and that's very ticklish to do both quietly and unobtrusively.

In the RGB LED system, you just look at the color of the light for each tube and diddle the bias pot until you get a green. In fact, on this setup, the biasing is an afterthought. It only cost the LEDs and pots to get it. The point was to sense the cathode currents and shut things down when there was a tube overcurrent that could cook a good tube, or fry a power or output transformer if a tube was shorted.

But like all designs, creeping feature-ism reared its head.

[martin manning]

Ah, I see. As I recall this thing I read about years ago would only adjust bias if it determined that the amp was quiescent for some predetermined time interval. Seemed too complex and costly to me at the time.

The RGB bias indicator thing would be pretty nice all by itself.

[R.G.]

We'll see how this works out. I actually posted that here to promote some thinking about what was and was not useful in something like this.

The $2.00 chip in the middle is of course a microcontroller. These days that $2.00 gets you a 28 pin chip with 11 10-bit A-D channels and a whole lot more that's simply unused in this design. It's being dramatically underused.

What I hand in mind was this:
- Powered by heater voltage, but in an odd way. I decided to use a small 120: 6V+6V transformer and use only the 6V secondaries. The 120V "primary" will be unused. That lets me use the heater power with no effect whatsoever on what else it's used for, as well as DC isolating the circuit for better sensing.
- A pushbutton that enables the LEDs for biasing. Press the button, the LEDs light up with the current too-hot/too-cold/just-right state of each tube, and can be tweaked. Let the button go and the biasing lights go out. The pots still change bias, but you can't see what you're doing...

When not biasing, the uC looks for overcurrents. It does this by reading the cathode currents that happen during playing. When this current goes over the level set by the "overcurrent" pot, the uC notes this and waits the amount of time indicated by the setting of the "time" pot, and if the over current goes away, forgets the overcurrent.

If the overcurrent persists for more than the preset time, it shuts down the B+ to the tubes, while remembering which tube(s) did the overcurrent. The Red LED for that tube(s) then blinks. A reset on the reset button resets the whole thing back to running B+, as does a power cycle.

So there are 1.2 plethora of lights and switches.

[teemuk]

http://www.tubesync.co.uk/

[R.G.]

Yep. Again, the automagic biasing of tubes.

Auto-biasing is OK, if overcomplicated, but there's much more that needs done than that.

I found the "be part of the advances as Tube Magic (or whatever) is integrated into major maker's amplifiers" to be a little amusing.

Really this is all pretty simple stuff.

[gui_tarzan]

The SVT CL I have on my bench right now has two bias adjust pots on the back with red and green LEDs. I don't know exactly how it can tell what voltage is correct (there must be a set voltage reading in a chip), but at idle when the bias is "right" they're green. As you're playing the red light comes on but the green light stays lit. It returns to just green at idle.

Granted it's not a self-adjusting bias, but it is kind of cool.

[R.G.]

Yep. I used a simple red/green LED on the Workhorse amps because RGB wasn't readily available and cheap back then. That's just about exactly what went in Workhorse.

Since it goes red during playing, I suspect they did it the same way my first protos did: run the voltage from the current sampling resistor into a set of comparators and let them drive the LEDs directly. I got "dancing LEDs" too.

The trick is to set the compare voltages to just below and just above the desired current. Between the two, light the green LED, and above, light the red one. There is some finesse involved in getting that "OK zone" small enough to be accurate enough, but large enough so that guitarists can actually hit it, not always be shooting through the green zone.

As I mentioned, I had a working autobias, but getting "green is good" was so fast and so easy that there was no point. Bias doesn't need tinkered continuously, no matter what magical tone sparkly-ness the advertsing copy says. If all biasing needs is looking at the back of the amp and diddling 2-4 trimpots, that can be done in under 30 seconds. It's so fast that you can re-bias between sets, even between songs if the singer will mess with the audience for a bit.

[R.G.]

I figured out something that had been bothering me - how to set the target bias current for a tube without building a meter into the bias/shutdown adapter.

If you change from, say, 6L6 to EL34 or vice versa, you will need to set a different target point. Or you may just like your 6L6s hotter or colder than normal - whatever normal is. So only if you set up the amp for a new type of tube or new preferred bias setting will the target current pot need calibration.

You just build in a pair of meter probe points in the adjustment panel. You plug your meter into the points and they connect directly to one of the tubes' cathode resistors. I'll call this the "master tube", just because its bias is set first.

You read the current in the external meter and adjust the bias pot for that tube to be the right current. The LED is to be ignored during this, and it may be any of the three colors.

Once the current in one tube is correct by meter measurement, you then
leave the bias pot for the master tube alone and go adjust the target current pot until the LED for the "master tube" is green.

This sets the indication current for all of the tubes.

The other three tubes can then be set to read green, and they will match the master tube's current within the tolerance band of the measurement.

With the target current set, you're good to go for any future tube swapping of the same type tube and you just set them to show green by adjusting their individual bias pots.

I'm still working on how to set the overcurrent shutdown pot.

[Gaz]

R.G. Just wanted to say this sounds really cool!

[Structo]

Yep, lots of great stuff there.

R.G.'s website has frequently been referred to here as a great resource.

[R.G.]

Thank you. You are very kind.

I did enough of a circuit, layout and physical design to get an early view of how this would work. You can see it here:
http://geofex.com/Article_Folders/Tube_ ... 0Panel.pdf

Drawing up pretty pictures of operator panels is all well and good, but this one was generated based on an actual PCB layout behind it, so it's fairly real. The page shows the panel real size, which is 1.5" by 4". There are two PCBs behind that, one with the switches, LEDs and controls on it and the other with the controller circuit, power supply and wiring to the rest of the amp.

As I mentioned, the circuit is powered by 6.3Vac from the heaters, through an isolation transformer. A suitable transformer can be bought new for $5 to $12. Well, I guess you could spend a lot more, but it can be as low as that.

The wiring to the amp is through a terminal block that you stick wires into and screw down to clamp.

I'm still thinking about how to mount this in an amp. There are conflicting requirements. It needs to be in the amp chassis to keep the critical bias wires safe inside the chassis, but the panel needs to be accessible for adjustments. I thought of something like hanging the panel from the lower side of a Fender combo chassis. The mounting is the most ticklish part of this.

There may have to be several versions to make mounting easy.

I have now figured out why the autobias adapters are so popular in the amp add-on community - it sidesteps the issue of a control panel entirely and buries the whole mess inside the amp. That may not be (and IMHO is not) the best way to do it, but you can hide it all inside and not deal with either the user interface issues or how to mount the blasted thing.

[martin manning]

R.G., where do you see this design going? An aftermarket bias gizmo? Might be hard to retrofit, but OEM's and DIY'ers would probably buy it. You could call it the Goldilocks... (but there's no running away at the end).

[R.G.]

Where is it going??

Oh, sure, ask pertinent, realistic questions about real world applications!! Where's the fun in that?


I have no clue where it's going. I just can't stop designing things. It's a personal failing.

Most likely, the idea will get some play with DIYers, then be largely ignored and forgotten until someone decides to put it in their production amps, and then just does so, crowing loudly in the advertising about how clever and original it is.

Given how very efficient it is, I wouldn't be surprised to see a patent application or two filed on it. I personally consider it to be something that's "obvious to one skilled in the art" and hence not patentable, but then I have had things that went up on geofex turn up in patents being granted before. And there's always Randall...

Mostly I think it's a neat idea and ought to be done somewhere. I already did the precursor in the Workhorse amps, minus the emergency shutdown functions. I did the discrete version of the emergency shutdown back in about 2000, and you see how far that got me.