Transistor equivalent

Status
Not open for further replies.

That's wise words from someone who knows. There are a heap of opamps that will outperform any transistor circuit you could hope to develop. Just make sure you do a decent layout and decouple the supply lines physically near the opamp, unless you want to have an oscillator instead of an amplifier that is. For years the NE5534 was the standard chip for audio work- low noise, low distortion and high drive capability and not expensive. They still sounds ok today but if you really want low distortion and a nice sound the modern range of audiophile opamps are out of this world. All opamps and other components, especially capacitors, sound different- you cant simply go by the data sheet. Whatever you do never put ceramic capacitors anywhere near your audio designs- just try it! They are ideal for decoupling in digital circuits and power supplies though. In that type of circuit you will see them sprinkled everywhere.

Just to let you know, the sallen and key filter is reasonably easy to build and get working ok, unlike some of the other filters. You do know that filters in the audio path are frowned upon these days!

You have got me going here, only because you remind me of the way I used to experiment with electronics. So can I give you one essential piece of advice. Always try to develop on thing at a time- in your case either Sallen and Key filters or amplifiers. That way you will have better results and more importantly- more fun.

I will get my coat now!
 
Last edited:
Yes all diodes can be used as thermometers, which is the essential part of modern disposable digital thermometers used in hospitals.

You can even heat up a power LED and measure the voltage drop with rising temperature or pulse it off and measure the Vf at 1mA to eliminate the variation in all LEDs due to ESR only. SO one can accurately measure the junction temp when they are driving them hard. Some are -3~4mV/deg C but depends on chemistry and can be found on most datasheets. ( but not the 1n4148)

FWIW there is a smart power diode designed for PV bypass solutions which has a Vf of 34mV @8A so at 10uA it will be close to zero. It has lower noise and Vf than Schottky. https://www.ti.com/lit/ds/symlink/sm74611.pdf

A full wave boost or drop of 3dB is hardly noticeable in volume in practice but a 3dB boost in sensitivity by lower detection threshold means significantly farther distance can be received. So going from 100mV to 1mV is an improvement of 20dB in RF power path loss gain since NF or noise factor is not the gating item in this simple detector.

The same for LED optical power or Luminous Intensity in mcd because our eyes and ears have a logarithmic sensitivity range of well over 100 dB when young.
 

Hello Bob,

Apologies not needed. I make sweeping statements often to simplify the issue. Right off I said that I have no experience with xtal radios, so I will be learning from someone who has. The other thing is that this is just a fun topic.

May I ask you if we are going to discuss this further not to use statements like ...false ...false ... misleading ... the correct theory... That is innuendo and circular, which doesn't bother me too much, but the problem is it is not constructive. Very often when people have different views it is because of a basic difference of interpretation at kick off. If you want to see an example of that just bring up the subject of matching.

After that, back to detectors. Can I take your points in turn:

(1) Semiconductor diodes start conducting at zero volts.

I have never seen one of those, and besides which it seems to contravene the band gap physics. Could you show a graph of Vf/If for the diodes you have in mind. Having said that, if there is such a diode in your field of speciality, it would be great to know more because it would open up all sorts of possibilities in circuit design- I mean that. Are you thinking of leakage current?

(2) the current to voltage relationship is a power law.

I don't understand the term 'power law' in relation to this thread- do you mean the exponential relationship of the Ebers–Moll model? Because, if so that clearly gives a point that could be considered a turn-on. Also I said quite early on that the xtal radio performance could be much improved by matching the antenna signal to the Vf/If diode characteristics, and suggested the use of a matching transformer. Or do you mean power as in watts? I have seen that in micro wave literature

(3) The forward voltage Vf (also ... referred to as turn-on voltage) is simply the voltage drop across the diode at a specified current (often 1 mA).

This is one of those basic kick off misinterpretations. Vf is the forward voltage of a diode. The turn-on voltage is one particular value of Vf. They are not the same thing. In fact, I said, in view of the experiments the guys have carried out, Vf seems to be the important characteristic for diodes in Xtal radios. I do not know why you mention 1 mA unless you mean the point on the diodes If/Vf curve where the diode is modeled and subsequently used as a microwave detector.

That may be the way that detectors are specified in detector diode specifications but that is just one point on the Vf/If graph. If I remember correctly, the diodes are biased with 1mA dc to get them beyond the cut-off point as I call it and also to ensure that they are operating at the point that the manufacturers have characterised. But in the case of microwaves a whole load of other characteristics are dependent on bias current. Almost none of them kick in at the frequencies being discussed in this thread. Also xtal radios don't have anything like 1mA If. The whole discussion on this thread relates to weak signal detection. See the Vf/If graphs that I posted previously.

(4) So, relying on Vf values is misleading.

That is to be established.

(5) Many weak radio signals are less than the typical Vf of a diode, but the diode will still detect the signal.

I am being serious here: what is a weak signal? How do you relate a weak signal to the Vf/If characteristics of a diode. What is the V/I characteristics if the weak signal. I think by weak signal you mean low power. I suggest that if you had a signal with a power of 1 mW with a V/A characteristic (Z) of 10V at 100uA almost any diode would do a fine job. But if the V/A characteristic were 0.1V at 10mA there would be nothing from the detector.

Also see my bit about bias above. Also see my previous comments.

(6) While Vf does give a ... indication as to whether a diode may be suitable as a detector, there are other more important parameters that determine how well it will work for weak RF signals, (signal levels well below Vf).

Could you let us know what those parameters are, as they relate to the relatively low frequency of AM commercial radio. I used to work in a microwave lab so I may have an idea what you are thinking about. If what I think is true you are talking about diode models at microwave frequencies- a very complicated area. But I have used microwave diodes for other purposes; they are just diodes, although dammed fast. GaAs diodes have an even larger band gap than silicon, but they have advantages that outweigh this.

(7) There are much better schottky diodes made for detector use, but the best one will depend on the characteristics of the rest of the circuit. It's an impedance matching issue. To get the most audio out, the detector has to be impedance matched to both the RF tuned circuit and the audio output device.

That is pretty much what I said. My whole object was to investigate which diodes are best.

When you say, 'audio output device', I'm taking it that you mean detector or do you mean headphones where the Z of the phones directly loads the detector in a simple xtal radio: 100K for Willen's xtal phones compared to the 32 Ohms for his Hi Fi phones. The first would hardly load the detector output, that's why they are used, but the second would completely short it out.

The 1N34A is still just about the best there is for the typical crystal set, and they can still be found with modest effort

How do you know this?
____________________________________________________________________________________________

I would be interested to hear what you think so please come back especially as you have sparked a thought; it could be that microwave diodes would make excellent xtal radio detectors. The only thing is they are hellish expensive, and extremely delicate. I bet you one of my 'like' awards that when we boil this down we are both talking about the same thing.
 
Last edited:
It's just occurred to me that if one of the emulation experts could do a run with a diode connected to a resistor, say 10K Ohms, with a sine wave input and plot against output that would be very interesting. Say start at 0V peak to peak and work up in 10 mV steps. I don't have an emulator myself or I would have a go.
 
2n3904 low noise when used in preamp? I have some opamp but they are lm358 and lm324 , only 2 ne5532

LM358 and LM324 are usefull devices and could be used to make a Sallen & Key filter, no problem. But if I understand you correctly you are hoping for hi Fi and they do not sound that good. The NE5532 is not far off the NE5534- same family- and would sound pretty good, as would the TIL071 and TIL084 families, both as cheap as chips. If you have any other op amps let us know and somebody will let you know if they are ok for audio work or not. Just to let you know, the two op amp families I just mentioned have a nice smooth sound and they are also very friendly to use, no hooting or other nasties. They are FET input so have essentially zero input current, which makes the biasing easier. THe NE5532/NE5534 is not brilliant in that respect. Do you have an electronic supplier where you are?

update 14 November 14
TIL07xx data sheet (they were launched in 1978 but are still inproduction- they are much better for audio work than TILo8xxx)
https://www.ti.com/lit/ds/symlink/tl071.pdf

You say you have 2x NE5532. They are twin amps in one pack so you have four amps total. Problem solved
https://www.ti.com/lit/ds/symlink/ne5532.pdf
 
Last edited:
2n3904 low noise when used in preamp? I have some opamp but they are lm358 and lm324
Simply read the datasheet: 2N3904 at the voltage, current, source impedance and frequency band listed, its noise is typically 5dB but no maximum is given, it might be 20dB which is very bad. A BC549 has maximum noise that is less at 4dB, its typical noise is much less which is very good.

Here is the horrible crossover distortion (it sounds like buzzing) of an LM324 or LM358 opamp with a sinewave as its input:
 

Attachments

  • LM324 crossover distortion4.PNG
    52.3 KB · Views: 411
In a practical sense the dc leakage current, and ac noise currents prevent detection at 0V hence a useful forward voltage and current depends on the Noise BW or Q of the front end filter.


I recall designing this single transistor amplifier with over 6o dB of gain BPF at 1MHz for a 50 Ohm loop antenna with an LC band pass filter with very low distortion using negative feedback. using a few mA and running off 1.5V

Here I simulate it with 10 uV of 1MHz sweeping from 0.9 to 1.1MHz you can change the sweep gen to 1m and see how much different V+max out vs V-max for the peak compression difference which is an accurate indicator of THD.

Normally with common emitter you avoid letting Vce get less than 2V for good linearity but with high Q and lots of negative feedback, one can reduce distortion considerable near saturation for Vce =between 0.5 and 2V as shown above.

I think the crossover distortion is reduced considerably with 2k pullup to 5V in the low current LM358 type.(at the expense of DC power wasted)
 
Spec,
If my tone was a bit confrontational, it was because I tend to get frustrated when people use diode forward voltage to explain low level detector behaviour, and for that I apologize. In fact, a lot of experienced crystal set builders insist on doing this, when they should know better.

In essence, the VF parameter that is typically given in a data sheet is the diode voltage drop at a specified current, usually 1 mA. This is a useful figure in large signal applications, but not so much when dealing with the extremely weak signals encountered in crystal radios. We tend to be dealing with microamps in a crystal set, so parameters measured at currents a thousand times higher tend to become irrelevent. The website, that I posted a link to, will give a much more detailed explanation than I can give here. But it is worth noting that on that page, he recommends making diode measurements at about 1µA.

However, to address your comments...

I'm sure you have, but maybe didn't realize. A typical 1N4148 behaves this way, but IV curves on typical datasheets tend to emphasize high level signals, and the low voltage behaviour is hard to read. However, have a look at the attached sample IV curves. You'll note that even with the voltage below the so-called VF, there is still current flow.

The last IV graph is an HSMS-2850 replotted as a log-log graph to highlight the fact that there is no point where the current abruptly stops flowing while the voltage is still greater than zero.
I guess it would have been more correct to say that it is an exponential relationship rather than power law. I was referring to the Shockley diode equation:
[latex]I=I_S\left( e^{\frac{V}{nV_T}}-1 \right)[/latex]
If you analyze the equation, you'll see there is nothing that can make the current drop to zero before the voltage does.

I'm not sure what you mean by cut-off point. I seems that you are still arguing that there is some positive voltage below which there is no current flow. However I refer back to the original quotation of yours which set me off in the first place:
As has been implied by the other posts, the forward voltage (VF) of the detector diode defines the lowest amplitude signal that will be detected. This will not be a gradual thing: below the detector diode VF you will get nothing.
Here, you seem to be implying that VF is a fixed parameter below which there will be no detection. I don't know how you could interpret this statement any differently.
But, this is getting into semantics, and we could discuss definitions forever. So I'll move on to other items.

(4) So, relying on Vf values is misleading.

That is to be established.
Fair enough. I've given a link to a very good source of information. But if you read the referenced information, you'll see that the author never even mentions things like forward voltage or cut-off etc. He is mainly concerned with saturation current, Is, which is the most influential parameter in determining how well a diode works as a detector. The ideality factor n is also important, but since this parameter doesn't vary much from one diode to the next, it of lesser concern than Is.

By weak signals I mean power levels of less than 10 picowatts, which are still intelligible with a decently designed crystal set.

How do I relate these signal levels to VF/IF? I don't. You were the one referring to VF, I wasn't.
Referring again to the linked site, the author goes into great detail about audio power out versus RF power in, based on diode impedance matching and the Is parameter and axis crossing resistance (which is derived from Is).

Primarily saturation current Is, and ideality factor n, as mentioned previously.
(Edit: These are static characteristics that have no reactive elements. So they apply from DC to normal radio frequencies. There are additional reactive elements that can be added to the model for dealing with microwave frequencies, but are not necessary for this discussion.)
Fair enough. I wasn't disputing what you had said, just restating it.

When you say, 'audio output device', I'm taking it that you mean detector or do you mean headphones where the Z of the phones directly loads the detector in a simple xtal radio
I was referring to the headphones, or combination of audio output transformer and headphones.
 
Last edited:
Bob,

I too have things that get me going- in general, anything that gets in the way of the engineering so I was a bit confrontional and I appologise for that.

What an excellent reply: as I have stated, I'm no expert in xtal radio detectors so what you say is very intresting. Maybe I'll manage to get to the freshman level one day. I'm a bit busy at the moment but I thought I would say hello. It will take me some time to understand the the points you have made but, in the meantime, I will still post on EOL for the simpler aspects which are quicker to do.

PS later:
Although I worked in a microwave lab, all the clever stuff was done by the front end gurus. They did the antennas, wave guides, GaAs amps, down conversion and detection and presented us mortals with a nice big signal at 50 ohms, complete with noise jitter etc, to process. They used to think it strange when we were discussing rise times of a nano second or so- to them that was just DC. It took me a while to dicover what Skolnik (Merrill Skolnik) and Watkins (Watkins-Johnson) were.
 
Last edited:

Hi again,

Yeah i was really thinking in terms of the commercial AM band which only goes up a little over 1MHz if i remember right. That means we might be able to use a synchronous rectifier and basically design for a variable conduction threshold. Would be interesting to see how this works out.

I meant to mention the bulk resistance too, which is also interesting.
 
Here is the horrible crossover distortion (it sounds like buzzing) of an LM324 or LM358 opamp with a sinewave as its input:

Thats some cross-over distortion. Of, course you can always bias the output into class A with a pull up (or pull down resistor) as has been said, but even so they still don't sound good.

I don't know if you remember the µA709. There was something very odd about that chip soundwise and I never found out what. The µA741 was a lot better. The µA741 and µA748 were used in the Texan amplifier as published in Practical Wireless in the UK. Don't know if you have heard of it. You could buy all the parts including PCB chassis and nice wooden case. Thousands wer built and I used to think my tweaked version was the last word. After all, an amplifier doesent do much and far exceeds the audio frequency spectrum and the distortion is only 0.1% etc ect. Then my son got into Hi Fi and bought a Cyrus Amp, Epos speakers, and Linn deck with moving coil cartridge. My tresured Texan never sounded the same again after hearing that. Having an engineering vieiw I have been proved wrong so many times in audio; yes the cartridge and spearkers may make a difference but speaker wire, loudspeaker stands, turntables no no.

Audio is engineering but it is also a black art. I ripped off the front end of the Cyrus but however much I tried it didn't have the same dynamics of the real thing.
 
Last edited:

I'm not sure what the upper frequency is for AM, as you say 1Mz down is where most of the commercial stations are, but I think there is some short wave stuff up to 30Mhz but not checked. I know a bit about synchronus rectification for SMPS but am not familiar whith any radio techniques where I could imagine you need a bit of finess rather thhan just blood and thunder.
 

Be just the sort of amp to go on the front of the xtal radio. Once the signal is amplified to a decent level things would be simpler. The only trouble is where do we stop- superhet next! That would be fun and probably not that hard to do simply.

Great Sim Tony, but you must keep things simple for me

Could I be cheeky and ask you to have a go at the model in the attachment, when you have the time that is- I see from the number of posts on ETO that you are pretty busy:

(A static graph of X_OP against X_IP would be ideal)

About sim packages, what would you recommend in the low-cost/free category?

The sims on EON are very useful, although a few seem to suffer from GIGO syndrome, especially from those just starting in electronics.

I have never done a run myself- at work there was always someone who would be only too pleased to do it for me, providing that I gave them the sim model that is. I think mainly P Spice and MatLab. I can't find any recommendations on EON for good packages but didn't want to post the question because I'm sure it's one of those evergreens that keep popping up. There is info on the net but I would rather get the opinion of the men at the front line.

 
Last edited:
In a practical sense the dc leakage current, and ac noise currents prevent detection at 0V hence a useful forward voltage and current depends on the Noise BW or Q of the front end filter.

As a favour to Spec...

Here is output of a triangle voltage input is shown on load R as V=I*R
.. neglecting leakage R and diode capacitance and leakage R and noise current , both which limit how high your load resistance can be , hence sensitivity of diode is much less than typical Vf at rated current but much greater than zero. Perhaps 5% of Vf rated@1A is a useful Rule of thumb in 10Meg load range... This affects sensitivity

you can edit any value ( right click) on schematic

or add any node to scope

FALSTAD SIM uses Java may be slow . Stop when not in use.
 
Last edited:


Many thanks for the sim Tony- what a powerful tool. It does show the 'cut-off point' as I call it as can be seen in the attached image.

Having said that, I agree with the points that Bob made in his second post but more about that when I can have a good look at the theory in the reference.

As you say, the leakage and noise limits how small a signal can be detected, at least that's how I see it anyway.

 
Last edited:
In fact it's a pull down, Tony. Checked the datasheet.
A pull up or a pull down resistor will eliminate crossover distortion but since the LM324 or LM358 can source twice as much output current as it can sink then the opamp can drive a lower load resistance if the added resistor is a pull down.
This added resistor must have its value calculated so that the output transistor stays in class-A for the minimum load resistance and entire output level.
 
Status
Not open for further replies.
Cookies are required to use this site. You must accept them to continue using the site. Learn more…