high side current monitoring

[R.G.]

As an eternal circuits kibitzer:
- you want matching of the first transistor Vbe and the diode. I'd use another same-kind transistor diode connected for the diode.

- you want to think about current gains; the currents divide in the ratio of the 10R and the 1K, I think. So the current in the 1K is 0-1ma. That means the base current should be 0-1ma/HFE. Transistor current gain changes with collector/emitter current, going down at low currents. This is the basis of most VCAs. I think you'll get linearity errors at low current. I don't know how bad that will be, but the concept has bitten me before.

- you want to think about divider-string currents. You'd like the base-emitter of the top transistor to have a similar forward voltage to the diode. Using a transistor diode connected for the diode helps, as the base-emitter of the diode-connected transistor/diode only conducts a similar current to Q1 and the collector-emitter conducts the rest. If you could make the collector current of the diode/transistor be the same as Q1, the Vbes would track as well as they could, and you'd only have the differences between them as offset/linearity errors.

That says that the divider string needs to conduct currents equal to the Q1/Q2 string. That's not possible with resistor loads, but some thought about the resistor values is in order to keep them similar and not starve or overload the resistor string.

Hmm. I wonder if the best thing to do is to use a double transistor ladder, mirroring Q1/Q1 on the R3/R4 side. That might even get you tracking and make HFE funnies go away.

[J M Fahey]

Can I say "Love this place"?
Or that phrase was already registered by Randall Smith at the US Patent Office?

[defaced]

My favourite circuit so far is the LT6101. If you had trouble sourcing the chip, you could easily reproduce it with an op-amp and a P-channel MOSFET. (You'll be needing a bunch of them for the cascode string as described by RG anyway. )

Look for an op-amp whose input common mode range includes the positive rail, like the TL0xx series. I forget which one is the low-power one, but you'll want that one, as all the power for it is dropped from the B+ supply.

The Hall effect current sensors ("LEM coils") are great for welding machines and 50hp inverter drives, but they're not much use for low currents.

Thanks for the input Steve. For posterity, the coil on the servo drive I mentioned is a LEM HY5-P, nominal current rating of 5 amps. So as I suspected, it would probably be ok for B+ in a large amp, but too large for screen current.

[R.G.]

As an eternal circuits kibitzer:
- you want matching of the first transistor Vbe and the diode. I'd use another same-kind transistor diode connected for the diode.

Yep. Works better in simulation.

- you want to think about current gains; the currents divide in the ratio of the 10R and the 1K, I think. So the current in the 1K is 0-1ma. That means the base current should be 0-1ma/HFE. Transistor current gain changes with collector/emitter current, going down at low currents. This is the basis of most VCAs. I think you'll get linearity errors at low current. I don't know how bad that will be, but the concept has bitten me before.

- you want to think about divider-string currents. You'd like the base-emitter of the top transistor to have a similar forward voltage to the diode. Using a transistor diode connected for the diode helps, as the base-emitter of the diode-connected transistor/diode only conducts a similar current to Q1 and the collector-emitter conducts the rest. If you could make the collector current of the diode/transistor be the same as Q1, the Vbes would track as well as they could, and you'd only have the differences between them as offset/linearity errors.

Yep. Changing hfe and resistor divider chains causes variable mismatching and offsets, etc. Using another transistor diode connected (i.e. base to collector) is more accurate.

That says that the divider string needs to conduct currents equal to the Q1/Q2 string. That's not possible with resistor loads, but some thought about the resistor values is in order to keep them similar and not starve or overload the resistor string.

Changing the resistor values in the divider changes things, in ways that are as yet mysterious to me. 100K is worse than 1M, for instance, at least with MPSA92 devices in the sim. Simulator didn't have MSB92s.

Hmm. I wonder if the best thing to do is to use a double transistor ladder, mirroring Q1/Q1 on the R3/R4 side. That might even get you tracking and make HFE funnies go away.

And we have a loser. Dual transistor ladder was worse on offsets and tracking than the original. No cigar.

Overall, accuracy varied from 0-maximum, being about 15% worst, 7% best, with some few "crossover" points where the measured and actuals coincided. I had to change the sense resistor to about 4K to get best tracking and scaling. Offsets were worst at 0ma, and ranged from 0.25V up to about 2V, depending on the whole slew of stuff.

My personal eval is that you could tinker it in for your setup if you had enough trimmer pots, or only cared about indications, not accuracy. Do the IC for accuracy.

[NateS]

On a roadtrip in the car on a tiny screen, which always means I miss stuff,.

Posting circuits on a discussion forum is pretty much invitation for kibbitzing.

I had an idea related to temperature alarm, that if it worked might actually be patentable. They really push patents at work. I'm a software guy, but still first patent is a nice chunk of change. But even if patentable, proving infringement would basically be looking for needles in competitor's haystacks. Then it occured to me, that they compensate for patents because you're assigning them the patent, and since my employer isn't in the tube amp power supply controller business, I think I'd technically have to license it from them.

When I get home I think I might put peanut butter on both sides of my sandwhich just to protest Schmuckers and the USPTO Patented until challenged in court application process.

Thanks for the detailed analysis and discussion. I'e certainly learned a lot from all participants. I

[NateS]

Hmm. I wonder if the best thing to do is to use a double transistor ladder, mirroring Q1/Q1 on the R3/R4 side. That might even get you tracking and make HFE funnies go away.

You mean something like this:


I decided to take your last bit of advice, however as I reconsidered the device (vs rolling my own HSCM) I dug into the datasheets and app notes, your warnings are in fact relevant to the device as well.



On the screen monitor, I can tolerate huge voltage drops, but sucking another 10mA off B+ is a steep price to pay. So there, I need to limit Vsense to keep Iout under control (~1mA)

On the heater circuit, where 10mA is nothing, I'm trying to watch my Voltage drop carefully to end up around 6V (-5%), both on purpose (-5% for tube life), and because that looks as good as I can get for as much headache as I'm willing to deal with - between the voltage drop on the current sense resistor, and the Rds of the MOSFET controlling heater current. (Originally it was just shorting out a chasis mount BFR inrush limit resistor. I decided that by swapping the BFR for a heat sink, I can keep it all on the board and get programmable current limit there too (for soft start).

I think the tempation to DIY on this one was the extra work required to get it down to ground, thinking, if I have to do the rest of it, why not?
[Sorry for the huge pics, but I was trying to keep the text readable and maintain attribution]

[Gingertube]

Something to print and put in your "Usfull info file".
The LT Current monitoring app note. Includes LT6101 circuits.

http://cds.linear.com/docs/Application%20Note/an105.pdf

Cheers,
Ian

[Steve Conner]

For posterity, the coil on the servo drive I mentioned is a LEM HY5-P, nominal current rating of 5 amps.

Neat! I didn't know LEM made them so small and cheap. My only experience is with the 400 amp ones that cost about $50.

In an earlier post that mysteriously vanished, I commented on Nathan's Cost Minimization Rule, to say that you should think about costing your time too. I guess you could argue that time is free on hobby projects, but not every part of a big, complex project is fun. It also serves as a reality check to help you finally get that project finished.

In my case, I'm planning a project where using a $25 LEM current transducer instead of a $1 current sense resistor would save a lot of time and debugging hassles. It is a one-off, so I'm not worried about recurring costs.

[NateS]

There is time, and then there's education. If we billed my time, let alone everyone who's contributed to the thread... well if we counted that way there wouldn't be an internet. I did decide to use the ZXCT1009 and cascode it to ground (a problem I'd have with any non-isolated solution). But...

I learned about cascodes, and BJT operation in general - places where previously I'd say "Ahh just stick a FET in there so I won't have to worry about base current and whats driving it." And I exposed a number of people smarter than me to a new part that appears useful and cost effective. And the fact that we ended up deriving the part design (even though it was hiding in plain sight), showed me things I need to watch using the off the shelf part.

Now I'm excited, once I get this project done, to start thinking about using a triode + FET cascode stage, maybe even using one on input, and a pair for a push-pull pseudo-pentode. 1 gain stage, DC coupled cathode follower, LTP phase inverter, and a push-pull 'pentode' using only two 12AX7's.

[R.G.]

Now I'm excited, once I get this project done, to start thinking about using a triode + FET cascode stage, maybe even using one on input, .

... like Musicman amps.

[R.G.]

Now I'm excited, once I get this project done, to start thinking about using a triode + FET cascode stage, maybe even using one on input, .

... like Musicman amps.