Why JFET amp instead of Non-inv Op amp? Do they both not have high Imp?

Noise. Although that's debatable now... low noise OpAmps are readily available.

...Jim Thompson

Thanks for the reply.

I understand that gain will go down with higher frequencies in accordance with the gain bandwidth product. What about the input capacitance issue? Does this increase with increasing frequency? I would be using a non-inverting op-amp so there would not be any miller effect. Where does the extra capacitance come from?

Thanks Stephen

Thanks again for the detail response!

If I could trouble you guys once more...

I am building a preamp/amp for the 60khz WWVB signal. Where should I do my filtering... after the JFET preamp and before the second stage amp? Or should I filter last (after 3rd stage)?

In general, do you filter before or after your gain sections? Why?

Again, I really appreciate your help.

Stephen

John P> > Thanks for the reply.

miller

get

that

the

stages

Depends what you mean by 'RF'. The majority of opamps operate only up to a few MHz. Some are designed to operate faster, but they are the exception at present. There are, of course, many high quality monolithic RF amps, but in answer to your question, there are two reasons:

1. A typical Opamp doesn't have the bandwidth, nor at any frequency does it have:

1. The really low noise figure of RF JFETs

Cheers

PeteS

Could you recommend a good low noise RF JFET?

This configuration of opamp has a very high impedance at low frequencies. At RF, the opamp has little or no gain, and its input may act like a bigger capacitor than a single fet.

Could you recommend a good resource for crystal filter design?

Thanks Stephen

Joerg wrote:

stage

stage.

first

around:

which

FET

still

No, the capacitance does not go up with frequency. But RF fets can have lower input capacitance than op amps, in many cases. Common source fet amplifiers are also dogs in many cases, and cascode or common drain configurations or other combinations may be needed to get reasonable gain at RF.

Can you specify what RF frequency range you are talking about?

In general, a chain of amplifier stages (there are at least 3 in most op amps) with overall feedback will not have as high a gain bandwidth product than a series of individual device stages with no overall feedback at the signal frequency because of the accumulated phase shift. This is why op amps have internal compensation capacitors that provide a dominant pole of roll off to get the gain below 1 before the total phase shift reaches 180 degrees. A series of independent stages do not have this phase shift limitation.

Sandpaper was discovered when a German inadvertently rubbed his toilet paper against the edge of the seat, rather than his ass ;-)

...Jim Thompson

I know, I speak from experience... had a very raw ass after my first week in Germany ;-)

...Jim Thompson

You're asking from the wrong vantage point. You perceive op-amps as the end-all, so expect them to be everywhere.

In reality, op-amps are relatively new, and for much of the four decades or so that they were available in solid-state form, they were quite limited in bandwidth.

For those who were around before they got useful to RF bandwidths, the question would not be "why aren't op-amps used" but "is there any good reason to use an op-amp in this RF stage?".

If a jfet supplies enough gain, then there's no need for anything more. And realize that receiver design isn't about "as much gain as possible" but in putting those gain stages where they will fit the receiver. You don't want or need high gain at the first stage of a receiver, especially if you've not given thought to having filtering there to keep out the unwanted signals.

IMpedance isn't really the issue. Oh, forty years ago jfets were kind of nice because they had high impedance like tubes, so it was easier to grasp their design for people coming off tube design, unlike bipolars which suddenly required much change because they were low impedance devices. And jfets back then had some, perceived or real, advantages in handling strong signals.

BUt then it was realized that if one was using higher current in the RF stage, it could better handle strong signals, and bipolars were back in favor. And the topology of the circuits changed, so low impedance wasn't the bogey it had seemed before.

Op-amps are overkill, unless you need them. If you use a single stage, then you can bias it exactly the way you want it. If you are using a whole opamp, that goes away.

Michael

Hello Stephen,

LC filter before the first stage, crystal filter after the first stage.

LC before because any high amplitude noise that can saturate the first amp would render the whole receive chain useless. Plenty of this around: Switch mode supplies, TV horizontal stages, big noisy brush motors, local AM stations, arc welding gear and so on.

A crystal filter needs a well-controlled impedance on both sides which makes it tricky to place right between antenna and amp. The antenna isn't a very controlled impedance.

The old RF rule is to place filters as far to the front as possible.

BTW, regarding opamps: As Jim pointed out there are some really great ones. But in commercial designs cost is paramount. While a very good FET can be had for under 5 Cents suitable opamps often cost more than ten times that. And they'd use more power. At 60kHz you might be able to find a decent opamp though, that's not a high frequency. I would still use a transistor but then again I am so used to designing with them.

Regards, Joerg

I read in sci.electronics.design that Stephen wrote (in ) about 'Why JFET amp instead of Non-inv Op amp? Do they both not have high Imp?', on Tue, 12 Apr 2005:

You want some filtering BEFORE any gain, otherwise you are amplifying stuff you don't want. The tuned antenna is your filter, but it's not a very good filter. So you need some more later in the amplifying chain. Exactly where is the best place depends on several factors, but the principle holds - don't waste gain-stage headroom on stuff you don't want.

Hello Stephen,

Why does it have to be a JFET? MOSFETs are more popular these days. My staple is the BSS123. That one is around 5c in large quantities.

If you use JFETs be careful, most don't have any sort of protection. I used a lot of BF245 JFETs when I was young. Just one healthy sneeze could send it to Lalaland if it wasn't in the circuit yet.

Regards, Joerg

Hello Stephen,

Mine has always been the ARRL Handbook. Any edition should do, especially older ones. Hwoever, for a WWVB application I'd look at schematics or app notes of the relevant receiver chips. Usually you don't need more than two crystals and some receivers make do with one.

With respect to app notes I believe the Temic receiver chip series is now courtesy of Atmel, so that company might be a place to start.

If you want to look at hobby projects and can't find enough for WWVB I suggest to also look for "DCF77", the European counterpart. Filter architectures aren't that different because it operates just a little higher at 77.5kHz. Some papers might be in German but at least the schematics should help.

Regards, Joerg

Hello Fred,

The BF245 is a JFET. I have lost quite a few of them but never blew the gate of a BSS84 or BSS123, so far. But maybe that could be chalked up to the fact that I was younger when I designed with BF245. ESD is learned with experience and between my JFET days and my MOSFET days there are about 10 years ;-)

For a WWVB receiver I'd probably start out with bipolars. Depending on where the antenna is they might get the occasional sting when weather runs through and bipolars stomach that better. It's amazing, we had a few small funnel clouds touch down recently and this is California. Here in the mountains that was accompanied by impressive lightning. So I have my antenna on a pole that can be unlatched and tilted down for a preamp swap. That's better than what they see in the flatlands: Missing roofs and trees.

My wife says yes. Birds fly away in a panic, cats scurry up trees. Which is why I keep a stash of a German Kleenex type brand. No joke, they were advertized as "snort proof".

Regards, Joerg

Hello Jim,

When you buy toilet paper at one of those super discount stores over there, chances are it's post-consumer recycle and the surface is comparable to 250-grit.

Regards, Joerg

"Joerg" a écrit dans le message de news:VBT6e.2146$ snipped-for-privacy@newssvr13.news.prodigy.com...

Huh.

Are you sure that these were JFets and not MOSFets? I've used loads of them, and specially the BF245 and never had to complain about reliability. MOSFets requires much more care if not protected, and specially the RF ones 'cause they have very low capacitances.

Or maybe you have a damn powerful sneeze...

The input signal going into an op amp is divided between two transistors (one on each input) and this is more noisy than if it just went through one, using a single transistor can also enable a more optimum operating point.

Of course the transistors u buy singly are somewhat diferent to the ones inside op amps too. also high frequency feedback loops in op amps are harder to work with.

Colin. =^.^=