Maintaining a Vbe Multiplier's bias value
in [Design]
|
Prev: Cushman CE-24A
Next: TIA Photodiode Bootstrap at 10MHz
From: George Herold on 9 Feb 2010 14:35 On Feb 9, 5:39 am, Jon Kirwan <j...(a)infinitefactors.org> wrote: > On Mon, 8 Feb 2010 19:16:24 -0800 (PST), George Herold > > <ggher...(a)gmail.com> wrote: > >><snip> > >"I'm wondering about additional topology changes to improve > >the performance still more." > > >Hi Jon, I've been 'sorta' following your thread on s.e.basics. I > >wonder if you abandoned class A operation too early? Why not keep > >things linear evreywhere and avoid the dead band? So what if you > >need a bigger heat sink. Its certainly a lot simpler. > > >George H. > > Well, George... No, I've not abandoned it. Actually, it's my > hope to wind up building the amplifier and then operating it > (by hopefully choosing a design where that is possible) in > different modes for the learning experience of it. I hope > that is in the cards. I really do. > > But to make a sharp point on it, although it's probably just > an extreme case, I remember reading about a 10W amplifier, > single channel, dissipating 120W! Creeps me out. So I > definitely _want_ to consider other classes of operation. And > cripes, I want to learn, anyway. So why not keep my options > open? > > Jon " I remember reading about a 10W amplifier, > single channel, dissipating 120W! " It might have been here, http://www.passdiy.com/default.html I got to reading about amplifiers on the above site... Do in part to your interest. George H.
From: Bob Monsen on 9 Feb 2010 14:40 "Jon Kirwan" <jonk(a)infinitefactors.org> wrote in message news:jg91n5d684ru5imsq1cfcjpjd1vddg2b2l(a)4ax.com... > I think this fits in sci.electronics.design, not .basics. > > > Jon Sorry, I didn't read the entire message... However, if you want a stiff multiplier, use a TLV431 instead of a BJT. Somewhat more expensive, but it'll be VERY stiff. However, you don't really want to hold that value constant. You want the voltage to compensate for the temperature of the output transistors. You might be able to use a diode to track the temperature change, and then use that in the feedback loop to compensate the TLV431. A honking big capacitor, one that has very low impedance at your frequencies of interest, is probably the best idea I've seen on the thread. On a related note, there was an article in a recent EDN about a self biasing preamp which was kinda cool. Instead of trying to track the difference using diodes or a multiplier, it used a couple of transistors and an opamp to set the correct values at the bases of the pass transistors. It was so novel (at least to me) that I typed it into LTSpice. Here it is: Version 4 SHEET 1 948 680 WIRE -288 -304 -608 -304 WIRE -16 -304 -288 -304 WIRE 144 -304 -16 -304 WIRE 320 -304 144 -304 WIRE 512 -304 320 -304 WIRE -288 -272 -288 -304 WIRE -608 -240 -608 -304 WIRE 320 -240 320 -304 WIRE 320 -128 320 -160 WIRE 320 -128 240 -128 WIRE 512 -128 512 -304 WIRE 448 -80 384 -80 WIRE 240 -64 240 -128 WIRE -608 -16 -608 -160 WIRE -160 -16 -608 -16 WIRE -128 32 -464 32 WIRE 48 32 -48 32 WIRE 112 32 48 32 WIRE 512 32 512 -32 WIRE 512 32 192 32 WIRE 560 32 512 32 WIRE 656 32 624 32 WIRE 672 32 656 32 WIRE -464 96 -464 32 WIRE 48 96 48 32 WIRE 128 96 48 96 WIRE 240 96 240 0 WIRE 240 96 192 96 WIRE 320 96 320 -32 WIRE 512 96 512 32 WIRE 672 128 672 32 WIRE -608 144 -608 -16 WIRE -608 144 -704 144 WIRE 32 144 -96 144 WIRE 448 144 384 144 WIRE 144 160 144 -304 WIRE -464 176 -464 160 WIRE -16 176 -16 -304 WIRE 48 176 48 96 WIRE 112 176 48 176 WIRE -96 192 -96 144 WIRE -48 192 -96 192 WIRE 240 192 240 96 WIRE 240 192 176 192 WIRE 320 192 240 192 WIRE -464 208 -464 176 WIRE 32 208 32 144 WIRE 32 208 16 208 WIRE 112 208 32 208 WIRE -160 224 -160 -16 WIRE -48 224 -160 224 WIRE -608 240 -608 144 WIRE -704 256 -704 144 WIRE -704 352 -704 320 WIRE -608 352 -608 320 WIRE -608 352 -704 352 WIRE -464 352 -464 288 WIRE -464 352 -608 352 WIRE -288 352 -288 -192 WIRE -288 352 -464 352 WIRE -272 352 -288 352 WIRE -16 352 -16 240 WIRE -16 352 -272 352 WIRE 144 352 144 224 WIRE 144 352 -16 352 WIRE 512 352 512 192 WIRE 512 352 144 352 WIRE 672 352 672 208 WIRE 672 352 512 352 FLAG -272 352 0 FLAG 656 32 out FLAG -464 176 in SYMBOL npn 384 96 M0 SYMATTR InstName Q1 SYMATTR Value 2N3904 SYMBOL npn 448 -128 R0 SYMATTR InstName Q3 SYMATTR Value 2N3904 SYMBOL pnp 384 -32 R180 SYMATTR InstName Q4 SYMATTR Value 2N3906 SYMBOL pnp 448 192 M180 SYMATTR InstName Q5 SYMATTR Value 2N3906 SYMBOL voltage -288 -288 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value 12 SYMBOL cap 224 -64 R0 SYMATTR InstName C1 SYMATTR Value 10�F SYMBOL res 208 16 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 56 VTop 0 SYMATTR InstName R2 SYMATTR Value 1k SYMBOL res -32 16 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 56 VTop 0 SYMATTR InstName R4 SYMATTR Value 100 SYMBOL voltage -464 192 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V2 SYMATTR Value SINE(0 .2 1k) SYMBOL cap -480 96 R0 SYMATTR InstName C2 SYMATTR Value 10�F SYMBOL cap 624 16 R90 WINDOW 0 0 32 VBottom 0 WINDOW 3 32 32 VTop 0 SYMATTR InstName C3 SYMATTR Value 470� SYMBOL res 656 112 R0 SYMATTR InstName R5 SYMATTR Value 8 SYMBOL res 304 -256 R0 SYMATTR InstName R1 SYMATTR Value 1k5 SYMBOL Opamps\\LT6234 144 192 R0 SYMATTR InstName U1 SYMBOL res -624 224 R0 SYMATTR InstName R9 SYMATTR Value 1k SYMBOL cap 192 80 R90 WINDOW 0 0 32 VBottom 0 WINDOW 3 32 32 VTop 0 SYMATTR InstName C5 SYMATTR Value 10pF SYMBOL res -624 -256 R0 SYMATTR InstName R3 SYMATTR Value 1k SYMBOL Opamps\\LT6234 -16 208 R0 SYMATTR InstName U2 SYMBOL cap -720 256 R0 SYMATTR InstName C4 SYMATTR Value 1�F TEXT 552 -296 Left 0 !.tran 0 .1 0 1u TEXT 552 -256 Left 0 !.four 1k 10 v(out) TEXT 552 -216 Left 0 ;.noise V(out) V2 oct 1001 1 100k Regards, Bob Monsen
From: Jon Kirwan on 9 Feb 2010 15:25 On Tue, 09 Feb 2010 06:49:11 -0500, Bitrex <bitrex(a)de.lete.earthlink.net> wrote: >Jon Kirwan wrote: >> On Mon, 08 Feb 2010 22:17:31 -0800, John Larkin >> <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >> >>> On Mon, 08 Feb 2010 22:11:51 -0800, Jon Kirwan >>> <jonk(a)infinitefactors.org> wrote: >>> >>>> On Mon, 08 Feb 2010 20:49:24 -0800, John Larkin >>>> <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >>>> >>>>> On Mon, 08 Feb 2010 20:43:03 -0800, Jon Kirwan >>>>> <jonk(a)infinitefactors.org> wrote: >>>>> >>>>>> On Mon, 08 Feb 2010 17:54:13 -0800, John Larkin >>>>>> <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >>>>>> >>>>>>> Hang a big capacitor across it. >>>>>> Nice try. >>>>>> >>>>>> Jon >>>>> No, seriously, that solves a bunch of problems. >>>>> >>>>> John >>>> Which problems does a slew-dependent, C*dV/dt bypass current >>>> solve? >>>> >>>> Jon >>> A big cap across the biasing gadget keeps the voltage drop across it >>> fairly constant, of course. That nukes some of the problems you >>> referred to. More peak current is available to the output bases, for >>> example. >> >> What size cap would help with power supply ripple? Seems the >> dV/dt is so small that a fair sized cap would be required to >> make any difference. Similarly for low frequency amplified >> signal out of the VAS. When you say "big," maybe you mean >> it. >> >> Ban is suggesting global NFB from output back to input. >> You've said as much when you say to apply "lots of NFB." I >> don't doubt the sincerity of either of you and I'm certain it >> will do a lot. But right now I'm interested in seeing what >> can be done right on this local subcircuit and at LF as well >> as higher frequencies. Unless someone wants to walk me >> through the thinking towards the larger concepts. I'm good >> either way, as it's the learning that takes place I'm looking >> for. But without such guidance, I need to move along at the >> pace I can handle while guiding myself. >> >> Jon > >Hey Jon, I found a derivation of the input impedance of the two-resistor >/transistor Vbe multiplier you might be interested in looking at: > >http://paginas.fe.up.pt/~fff/eBook/MDA/Mult_Vbe.html That one takes an approach that I'm not familiar with and didn't take. I'll have to consider the approach more. However, I did take a look at the end of it. It says: R = (R1+(R2||re)) / (1+(1/R1+gm)*(R2||re)) If I understand the value gm, and I may not, it's just 1/re or else re=1/gm. Basically, just the (kT/q)/Ic I'd mentioned when I wrote. If that is the case, I used these to see how that page predicts: ic=.005 vt=k*300/q gm=ic/vt re=1/gm r1=1000 r2=1000 r2p=r2*re/(r2+re) and then computed: (r1+r2p)/(1+(1/r1+gm)*r2p) and got: 502.5719049 Ohms. This is so far from my own calculations of about 15.4 Ohms that I just _had_ to put this into LTspice and test it. To do that, I simply set up the basic circuit with the two resistors and BJT and then hooked up a variable current source to the topside. I set it up as an AC source of 5mA with peaks of 500uA, and then ran a .TRAN on it and plotted the upper rail of the structure's voltage. I used a 2N2222 BJT, as well. Convenient, and I have them laying about. Anyway, so I ran the sims and got 17.44mV, peak to peak. Divided by the peak to peak current variation of 1mA gives an apparent R of 17.44 Ohms. My calculations arrived at 15.4 Ohms, or so. All this could be operator error. I may be operating the web page you suggested incorrectly, so that the 503 Ohms I get is because I didn't know what I was plugging in and where. I may be operating LTspice incorrectly, so that it's results aren't usable and it's just luck that the numbers worked out in my favor. But there it is. Here is the LTspice file: Version 4 SHEET 1 880 680 WIRE 128 0 16 0 WIRE 224 0 128 0 WIRE 288 0 224 0 WIRE 128 32 128 0 WIRE 16 112 16 0 WIRE 224 112 224 0 WIRE 128 160 128 112 WIRE 160 160 128 160 WIRE 128 208 128 160 WIRE 16 224 16 192 WIRE 128 320 128 288 WIRE 224 320 224 208 WIRE 224 320 128 320 WIRE 128 336 128 320 FLAG 128 336 0 FLAG 288 0 V_rail FLAG 16 224 0 SYMBOL npn2 160 112 R0 SYMATTR InstName Q1 SYMATTR Value 2N2222 SYMBOL res 112 192 R0 SYMATTR InstName R1 SYMATTR Value 1k SYMBOL res 112 16 R0 SYMATTR InstName R2 SYMATTR Value 1k SYMBOL current 16 192 R180 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName I1 SYMATTR Value SINE(5m 500u 50) TEXT -76 296 Left 0 !.tran 1 >For bypassing purposes the rule of thumb I've always heard is to make >the impedance of the capacitor 1/10th the value of the impedance looking >in to the circuit at the lowest audio frequency. Well, let's assume that I got lucky and LTspice and I agree on the figure of about 16 Ohms. With a signal at 20Hz, we are talking: C = 1/(2 PI f (R_ac/10)) = 5000uF Yikes. John L. wasn't kidding when he wrote "big." Luckily, in steady state it could be a low voltage cap! Jon
From: Jon Kirwan on 9 Feb 2010 15:26 On Tue, 09 Feb 2010 07:03:18 -0800, John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >On Mon, 08 Feb 2010 22:43:19 -0800, Jon Kirwan ><jonk(a)infinitefactors.org> wrote: > >>On Mon, 08 Feb 2010 22:35:05 -0800, John Larkin >><jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >> >>>On Tue, 9 Feb 2010 00:28:27 -0600, "Tim Williams" >>><tmoranwms(a)charter.net> wrote: >>> >>>>"Jon Kirwan" <jonk(a)infinitefactors.org> wrote in message >>>>news:okr1n55h5dvjjklg760dllkqq50v7s38ib(a)4ax.com... >>>>> Part of the function of the Vbe multiplier is to also track >>>>> the Vbe requirements for the output stage as it heats up and >>>>> cools down. >>>> >>>>The general idea is to put the Vbe transistor on the same heatsink as the >>>>outputs, if not glued to a transistor directly. >>>> >>>>Unfortunately, for widely mismatched current densities, this doesn't work. >>>>http://webpages.charter.net/dawill/Images/Ampere.gif >>>>In this boringly typical circuit, the 2N3904 Vbe mult. doesn't have enough >>>>tempco to compensate the far beefier (= lower current density??) output >>>>darlingtons. >>>> >>>>I was thinking of adding another CCS so a constant voltage drop appears on >>>>the Vbe's base divider resistor. Algebraically subtracting a fairly stable >>>>voltage results in the effective tempco (percentwise) increasing. The base >>>>divider ratio has to be changed to compensate. >>>> >>>>> In this case, I want it to track the output stage so I'm >>>>> going to have to couple it thermally in some useful way. What >>>>> I'm considering, right now, is how to make it immune to >>>>> unregulated supply variations and VAS output voltage swings. >>>> >>>>Don't worry about stability -- as John said, bypass and forget about it. >>>>Most of the dynamic VAS/CCS current flows into the output stage, since >>>>that's what it's there for anyway. The capacitor helps turn on the N side / >>>>turn off the P side for rising edges and vice versa. >>>> >>>>As for PSRR, the CCS's and gobs of feedback keep that in check. Of course, >>>>in principle you need something to start the CCS's. ICs do this with a JFET >>>>(i.e. current regulating diode) or bandgap reference (e.g., TL431), or >>>>sometimes both, to set a master current, from which everything else is >>>>mirrored. Most discrete circuits just use a resistor, which is "0%" PSRR, >>>>but it's not all that bad because the currents are balanced (*on average*, >>>>which means you'll see IMD products when it's moving). >>>> >>>>Tim >>> >>>This topology, thermally coupled Vbe multiplier, was mediocre 50 years >>>ago. And still is. >> >>A trek of a thousand miles starts with but the first step. >> >>Jon > >A trek of 23,000 miles starts with but the first step in the wrong >direction. There is no wrong direction at the start. It's all good. Jon
From: Jon Kirwan on 9 Feb 2010 15:56
On Tue, 09 Feb 2010 08:03:14 -0700, Jim Thompson <To-Email-Use-The-Envelope-Icon(a)My-Web-Site.com> wrote: >(1) Split R1, bypass that junction to ground Understood. >(2) Make R5 and R6 into mirrors, resistor feed from VDD, but split and > bypassed. I had been thinking more like this structure: >: to input to voltage >: stage mirror amp stage >: | ,---, | >: | | | | >: | | | | >: | gnd | | >: | \ | >: | / R4 | >: | \ | >: | / | >: | | | >: | ,------+ | >: | | | | >: | --- C1 | | >: | --- | | >: | | \ | >: | | / R3 | >: | | \ | >: | | / | >: | Vdd | | >: Q2 c\| | R5 |/c Q3 >: |-----------+----/\/\-----| >: e<| | |>e >: | | | >: | R1 |/c Q1 | >: +---/\/\----| | >: | |>e \ >: | | / R6 >: | | \ >: \ | / >: / R2 | | >: \ | | >: / | | >: | | | >: Vdd Vdd Vdd However, I take your point. >(3) As you said, replace R4:R3 with a mirror, I don't think a compound >device mirror, such as Wilson, is necessary. Understood. Although I'm not able to make my own decisions on this, yet, I've read repeatedly that the distortions to deal with are not at the input stage. The input stage can be made better, the improvements are small in comparison to what remains in the rest of a well-designed system. Point taken. >Study this if you want more info.... > > http://analog-innovations.com/SED/EnhancedCurrentMirrors.pdf Thanks, I will! >(4) Since you're on a learning curve, just replace D1/D2 with 1.5*Vbe, >losing about 1/5 of the Q3 quiescent current in the resistors. Thanks for taking a moment to confirm the "1/5th" division. I'd already figured that was commonly done and had some ideas of my own about why that makes sense. (I could talk about that, but I'm sure you already know and I think I know, too.) >Bypassing base-to-base (of Q4-Q5) will help at all but very low >frequencies. Okay. That's how I see it, too. >(5) Long haul as you "oomph" the power: > Q3 goes to Darlington, as do Q4 and Q5; D1/D2 becomes more >complicated (Darlington extension of Vbe multiplier). This is what I'd like to explore, right now. Extensions. It's because it is where my mind is at, right now. And I want to explore this more fully before walking away from it and moving on. >Start simple, then grow it, that way you learn before you flame it ;-) hehe. Good advice, of course. As I'm still struggling to make sure I understand each piece, right now, I'm just not yet ready to put it all together -- not even in a low power system. I might be able to vaguely grasp what I would be doing, but I prefer taking each part and thoroughly looking at its function before moving on. Then, when I look once again at the whole, I can better "read" what I see and that helps a lot in terms of gaining a global view. I'm still in the trenches, right now, and not allowing myself to raise my head much above that until I get some of the details nailed down. Speaking of that, can you confirm (or correct) the equation I developed for the simple Vbe multiplier's small signal R? Or the relative scale and _sign_ of the Early effect correction to it, which I peg near -1 part per thousand in the case I cited? All this is good for me to go through. Thanks, Jon |