Atlas dca model:DCA55 any good?

[PeakAtlas]

Thanks for the accuracy spec. Whenever IDSS is >5mA it shortens VDS. I think that it would be useful to keep that parameter steady and independent.

I agree, I think that would make a nice enhancement. I'll discuss with my esteemed colleagues here. Thank you for your feedback.
Cheers,
Jez

[Salas]

Thanks for taking it to consideration. Examples even when on PC, attached. Data line of interest is bottom left for those unfamiliar.

[PeakAtlas]

Thanks for the screenshots. Am I right in saying that you'd like Vds to be a constant figure for the test when Vgs=0? That would yield a variable IDS that would change from part to part, is that what you'd like to be able to see to compare parts? I'm sure we could do that. Although there are some JFETS that want to sink lots of current for a VGS of 0 for even tiny VDS values.

By the way, you could perform a ID/VDS plot of your JFET with VGS set to 0 (or any other value for that matter). You can plot the same graph for any JFET you like and then compare them on the same graph (see attachment that I did for 3 samples of a 2SK170 audio JFET).

Cheers,
Jez

[Salas]

Yes. Id when Vgs=0 at a steady Vds value is the definition of Idss. Fundamental in characterizing and matching JFETs.
Vds is 10 or 15V specified in most datasheets. 10V is a good compromise even for high Idss JFETs I have seen. They tend to become only marginally stronger CCSs above 10V.
I am aware that I can go in tracer mode and set some parameters, but such a method is time consuming for when searching for matches through a batch. Identify mode with steady Vds option would be as quick as when matching HFE for BJTs.

[PeakAtlas]

OK, leave that with us, we'll see what we can do.

Thanks once again for your feedback,
Jez

[PA4TIM]

In answer to your H_FE accuracy question, the H_FE is measured to an accuracy of nominally ±3%.

I hope that helps,

Jez

Just for the record, it is h_FE and h_fe,  just two of the many " h" or hybrid parameters that excist.
The first is the DC amplification of a common emitter amplifier (F=forward, E= common emitter, capitals is DC ) and I think this is what the peak measures. The other one is the result of delta Ic to delta Ib, sometimes called  small signal amplification. At DC the amplification can be 100 but at some frequeny, called Ft,  the gain is dropped to 1 so if your peak shows you a gain of 100, this will not say this transistor has still a gain of 100 in your application at 10 MHz. But Datasheets show it so it is just a test. I have seen transistors with a perfect gain at a steady Ib, but an increase of Ib did not give a single reaction. This was while repairing an old HP. The Peak and some transistor testers told me it was good. The scope told me there was a problem. The Tek curvetracer showed me the real problem. But these are exeptions, just to tell you, never trust any instrument without thinking or cross checking.

h_fe is a bit more difficult to measure.
From the wiki: https://en.wikipedia.org/wiki/Bipolar_junction_transistor#h-parameter_model


Etymology of hFE[edit]
The 'h' refers to its being an h-parameter, a set of parameters named for their origin in a hybrid equivalent circuit model. 'F' is from forward current amplification also called the current gain. 'E' refers to the transistor operating in a common emitter (CE) configuration. Capital letters used in the subscript indicate that hFE refers to a direct current circuit.


http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=1446653&url=http://ieeexplore.ieee.org/iel5/5/31082/01446653.pdf%3Farnumber%3D1446653

Both the ESR60 and ESR70 have the same measurement resolution of 0.01 Ohms and this is measured at the "industry standard" of 100kHz that most capacitor manufacturers like to specify.

This is not complete true. Most manufactures state |Z| at 100kHz and ESR as a function of D at 1 kHz or 100/120 Hz. Z = impedance  = (R -jX) or ( ESR and reactance). reactance = 1/(2pifC)  and ESR = D x reactance

Because at 100kHz  the jX of Z is neglectible for large enough caps, ESR will be the main part of Z. My home made ESR meter measures down to 100 nF and even lower. A very good 1 nF cap with a low D of 0.001 has an ESR at 1 kHz of 159 Ohm. So that is why it is more convinient to measure at higher frequency. But you can not measure C with 100 kHz because the appearant capacitance there differs a lot from the true capacitance. That is why 100 Hz and 1 KHz will stay standards for measuring C

But if capacitance is dropped a lot, Z will be higher. An impedance meter shows you there is "some" problem. But not if it is ESR or capacitance. The peak will only show the ESR and that can in some rare cases (got some samples here) still be good . And because the peak measures C using DC it will show a high capacitance so it seems like that cap is still good. But again, for 95% of the cases it performs very good and is a nice help next to your standard scope probing for ripple curents.

I tested some power MOSFETs a while back in Ton and Toff and beside a function of Vgs this is also a function of Id and Vds. A 600V 20A Fet is a lot slower at 15V and 1A

[Salas]

OK, leave that with us, we'll see what we can do.

Thanks once again for your feedback,
Jez

Cheers. BTW congrats on DCA Pro, its a most versatile little beast.

[Salas]

In answer to your H_FE accuracy question, the H_FE is measured to an accuracy of nominally ±3%.

I hope that helps,

Jez

h_fe is a bit more difficult to measure.

I agree, its a multifaceted parameter.
On another matter, the one of Vbe reading both on DCA 55 and Pro, I have seen that Ib is chosen in the mA. That one helps to max Vbe. Alas, its not representative of small BJTs in average biasing when in circuit. It would be better for Ib to be selectable in identify mode or about 100uA. Or not? Was high Ib chosen so to encompass power transistors, Jez?

edit: Here is a 2N4403 by ROHM. It would show no more than 0.65V Vbe in most circuit applications within its dissipation ability.

[PA4TIM]

I noticed that too. I have an early Philips transistor tester that can source so much base current it kills a BC547 in a flash. So you must be carefull setting it up. But the same goes for a curvetracer.  The problem will be that all those features make it difficult to operate and most users probably have no clue about how or what to do. They want to know, what is it, does it still work and what pin is what. And the Peak is good in that case as it is..

[Salas]

DCA series is also much used for HFE characterizing & matching purposes with differential and symmetrical circuits in mind. For Vbe also, since it can be decisive for tempco in current mirrors, parallel low noise BJT arrays, and the like sensitive applications.

[Salas]

Forgot to mention: Current hogging in parallel BJT power output stages is a make or break situation and you would want to match Vbe for also keeping individual emitter degeneration resistors low.

[PeakAtlas]

I agree, its a multifaceted parameter.
On another matter, the one of Vbe reading both on DCA 55 and Pro, I have seen that Ib is chosen in the mA. That one helps to max Vbe. Alas, its not representative of small BJTs in average biasing when in circuit. It would be better for Ib to be selectable in identify mode or about 100uA. Or not? Was high Ib chosen so to encompass power transistors, Jez?

edit: Here is a 2N4403 by ROHM. It would show no more than 0.65V Vbe in most circuit applications within its dissipation ability.

On the subject of HFE (and hfe), the main measurement parameter for gain that the DCA55 and DCA75 show is for HFE (DC current gain). Using the DCA75 on the PC it is possible to derive small signal gain (hfe) by grabbing some Ic graphs for various base currents and calculating slopes at various points in Excel. That also allows you to set numerous bias levels to see what works for your desired application.

On the subject of the ESR60 and ESR70 mentioned in an earlier post, both of these instruments only measure the real part of the complex impedance of the capacitor, effectively ignoring the capacitor's reactance. That ensures that the reading of ESR is not distorted by the relatively large reactance of a 1uF for example. We have been very careful to ensure that the ESR measurement is accurate and extremely representative of the real part of the complex impedance. Believe it or not, there are meters out there that only measure the bulk impedance and that doesn't work well for low capacitances even at 100kHz.

Turning now to the question of the test current used to measure VBE. It's been a tricky one this. Yes, generally we've derived the test current of 5mA to be as universal as possible for getting a representative VBE measurement for a wide variety of transistors including power transistors and even darlingtons that have shunt resistors across their base-emitter junctions. A small current (of say 100uA) wouldn't overcome the low resistances that are sometimes present in some darlingtons.

The user guides for both the DCA55 and the DCA75 has some further information on the base test current used for the gain measurement and for the VBE measurement (gain measurement is generally at much lower base currents of course).

The DCA75 however does allow you to sweep the base-emitter current in the PN junction graph, that allows you to see what happens below and above the normal 5mA figure.

Generally though, the difference in VBE that you get from a base test current of 100uA or 5mA isn't likely to affect your choices of biasing components very significantly.

Regards,
Jez

[PeakAtlas]

DCA series is also much used for HFE characterizing & matching purposes with differential and symmetrical circuits in mind. For Vbe also, since it can be decisive for tempco in current mirrors, parallel low noise BJT arrays, and the like sensitive applications.

I agree, gain and VBE are good parameters for matching purposes. Especially for current mirrors or paralleling applications. I've found that, for a particular variety of transistor (say a 2N3055), if the VBE matches well at one current (say 5mA) then it continues to match well at other base currents too.
Jez

[PeakAtlas]

Forgot to mention: Current hogging in parallel BJT power output stages is a make or break situation and you would want to match Vbe for also keeping individual emitter degeneration resistors low.

Very good point. The improved VBE measurement resolution and the adjustable test current that you get on the DCA75 is particularly helpful for this. It's good to keep emitter resistors at a sensible low value for maximising gain, but I tend to play it safe and use a little more resistance than I can probably get away with, especially for applications where the transistors may get hot, VBE with it's negative tempco can lead to problems especially if transistors are unevenly heatsinked.
Jez

[Salas]

if the VBE matches well at one current (say 5mA) then it continues to match well at other base currents too.
Jez

Eventually, that would have been my next questioning post about, was already in preparation.
If the Vbe reading at 5mA Ib is a good indicator that two matching BJTs will generally match Vbe at other Ib current levels also, then its enough info for quick shifting through a batch.

[Salas]

It's good to keep emitter resistors at a sensible low value for maximising gain, but I tend to play it safe and use a little more resistance than I can probably get away with, especially for applications where the transistors may get hot, VBE with it's negative tempco can lead to problems especially if transistors are unevenly heatsinked.
Jez

Julian Vereker would have been chuffed with DCAs in the hands of his production line staff. Can make a pocket tester, but can't turn back time...Yet.