Hi, I've never found any information about the optimal voltage across the base and emitter of the transistor of a CCS.
Of course the voltage must be well below Vbe and the higher this voltage, the higher the sensing resistor can be, so hfe times the impedance of the CCS.
But what is the sweet spot for it?
Thank you in advance,
Kind Regards
Of course the voltage must be well below Vbe and the higher this voltage, the higher the sensing resistor can be, so hfe times the impedance of the CCS.
But what is the sweet spot for it?
Thank you in advance,
Kind Regards
So you are saying the optimal Vbe must be well below Vbe?
I think something is lost in translation here... Vbe depends on the emitter current and device temperature, and very slightly on the collector voltage (Early effect)
I should have added a schematic, indeed.
The zener (or led) from base to ground gives a reference voltage (once subtracting the 0,7V you mentioned).
Is there an optimal voltage for that zener (or led)?
Optimal voltage? Depends on your use. Voltage loss in Re is voltage swing you don’t have. Do you need or want to swing rail to rail? Are you using it on a VAS and want to use it to intentionally limit how close to the rail you get? Where that doesn’t matter, higher Re gives more degeneration, which result in higher output impedance and lower effective output capacitance without resorting to adding a bunch of transistors. The reason they do that in ICs is that matched transistors are free and resistors BIG and poor tolerance. For biasing a diff pair, soaking up a good 15 volts in the emitter resistor isn’t necessarily a bad thing. You’d never see an op amp biased that way, but no reason not to with discretes. With tubes you have lots of voltage on tap, and you might use a 68V zener.
Seems not to be in accordance with the usual equation
https://cr4.globalspec.com/thread/68055/Voltage-vs-Current
@wg_ski
I’m thinking mainly about loads for source/cathode followers and tails for phase splitters.
No need to swing rail to rail, just to keep the ac impedance high enough without adding another bjt in totem.
I know hfe is affected by temperature and current, but there could be other details I’m not aware of.
I’m thinking mainly about loads for source/cathode followers and tails for phase splitters.
No need to swing rail to rail, just to keep the ac impedance high enough without adding another bjt in totem.
I know hfe is affected by temperature and current, but there could be other details I’m not aware of.
With 15 to 68 volts worth of emitter degeneration you can safely IGNORE all the secondary effects, unless exact temperature compensation is needed. I’ve never needed or wanted it in a tube amp, for an SS diff pair you can make a case for keeping it from drifting.
Just don’t use a transistor with an hFE of 20 (Which can happen on HV transistors). Just something reasonable so you can maintain bias in the zener. MJE340’s are usually plenty high enough. Typical 160 V sustained-beta driver transistors work well too. You can always get a current production type somewhere, even if it want what you used last year. MPSA42’s are fine if you’re not dissipating the universe.
Just don’t use a transistor with an hFE of 20 (Which can happen on HV transistors). Just something reasonable so you can maintain bias in the zener. MJE340’s are usually plenty high enough. Typical 160 V sustained-beta driver transistors work well too. You can always get a current production type somewhere, even if it want what you used last year. MPSA42’s are fine if you’re not dissipating the universe.
Not only artosalo, none of us understands your question which does not make sense , even less without a schematic.
Now that you posted it:
1) you are not talking Vbe but voltage from Base to ground.
2) Vbe is "one diode" or around 0.7V, no matter the Zener diode voltage.
3) no matter what voltage you apply to Base, Voltage at emitter will rise or fall as much as needed to keep Vbe difference around 0.7V
Basic transistor Electronics.
4) Vbe is not a random value but a difference.
In this case between base voltage to ground and emitter voltage to ground.
And it self-regulates.
5) so much so, that one of my troubleshooting techniques for mystery amps with no schematic available is, precisely, to measure all Vbe.
About 1000000 tines faster and less damaging than random unsolder - pull - test - resolder.
For, say, an NPN transistor:
* I find 0.65Vbe?
Good.
* +3V?
Base open
* -4V?
Some other transistor is open or shorted and this one is desperately trying to return DC values to normal
And so on and on and on..
Thanks, indeed that value is not critical in tube applications.
Anyway I’ve seen the Vbe thermal coefficient is -2 mV/°C, and from this document ( https://www.microsemi.com/document-portal/doc_view/14628-zener-voltage-regulation-with-temperature ) it seems that zeners around 3,9 to 4,7 V naturally balance that value.
Some leds should balance as well.
I’be often used 2N5551 in totem pole ccs because they have a good hfe (around 150 between 1 and 20 mA at reasonable temperatures) with red leds from base to ground, but being currents usually very low (between 1 and 10-20 mA) so dissipations not high, I would like to use zeners and avoid the totem pole configuration.
With due respect, wg-sky did not answer
voltage across base and emitter?
Not a word about it.I did, and in abundance of detail.
WTF?
Your question reads, literally:
Vbe must be lower than Vbe.
Does that make sense?
FWIW Mark Tillotson(and any other reading this thread) read exactly the same as I did:
So you are saying the optimal Vbe must be well below Vbe?
No need to drop into the “what the ****” kind of words. Thanks. I’ve got the answers I needed.
The transistor datasheet usually includes a plot of Vbe versus Icollector at room temperature.
For example, the bipolar transistor which Nelson Pass often uses in his current sources is a DiodesInc / Zetex ZTX450. I've attached the relevant curve from its datasheet, below.
As you can see, Vbe varies with Icollector. (Vbe = 0.60V @ Icollector = 1mA) and also (Vbe = 0.67V @ Icollector = 10mA) and also (Vbe = 0.76V @ Icollector = 100mA)
So the thing to do, is (A) choose your transistor type; then (B) grab the Vbe-versus-Icollector curve from its datasheet; and finally (C) read off Vbe at your desired Icollector. And Bob's your uncle.
_
For example, the bipolar transistor which Nelson Pass often uses in his current sources is a DiodesInc / Zetex ZTX450. I've attached the relevant curve from its datasheet, below.
As you can see, Vbe varies with Icollector. (Vbe = 0.60V @ Icollector = 1mA) and also (Vbe = 0.67V @ Icollector = 10mA) and also (Vbe = 0.76V @ Icollector = 100mA)
So the thing to do, is (A) choose your transistor type; then (B) grab the Vbe-versus-Icollector curve from its datasheet; and finally (C) read off Vbe at your desired Icollector. And Bob's your uncle.
_
Attachments
Quoting this because I like it and agree 👍
For my small anode load CCS, I use Zetex ztx560, with BC560C (on top IIRC, cant find the sch image to upload, it may already be here somewhere), lots of gain on tap, which apparently helps.
It certainly worked better for me than a single device like 10M45 or 10M90, LND150 etc, though I tried many simple configurations. I find these small BJT circuits lasted, the FETS have their own difficulties
It seems to work well, but I am not half as talented and knowledgeable as the experts here.
The compliance voltage is the voltage the CCS has to work with. i.e. below this the CCS will not regulate. For your zener solution is the zener voltage +Vce sat (about .3V). Sometimes with an LPT stage you may not have much cathode voltage if the grids are biased at 0V. In this case a current mirror may offer a better solution as they can operate down to 1V.
MOSFETS and JFETS make poor CCS as the current you get varies considerably from one device to the next. You then need a pot to set it.
MOSFETS and JFETS make poor CCS as the current you get varies considerably from one device to the next. You then need a pot to set it.
For a current source, the greater the emitter resistor, the greater the voltage feedback to base emitter junction as the base is held relatively constant. This would imply using larger Vb, while remaining within practical limits as it takes a larger voltage. Above about 5.6V diode operation switches from Zener (Negative TempCo) to Avalanche (Positive TempCo), and this can help improve the circuit temperature stability. That said for a LTP at low current (< 10mA) it is probably irreverent. I generally stick to 3.6V zeners because I have a bunch.