Page 1 of 1

BC109C Replacement?

PostPosted: Mon May 18, 2015 8:16 am
by Lawnboy
What is the best modern equivalent of the BC109C? They are getting harder and harder to find and surely there must be an improved version available now. I'm using them in a virtual earth photocell preamp, so low noise is important.

-Byron

Re: BC109C Replacement?

PostPosted: Mon May 18, 2015 6:04 pm
by Harry Dalek
i have always just used the BC547 548 549

Re: BC109C Replacement?

PostPosted: Mon May 18, 2015 7:56 pm
by Steve Anderson
The BC109C is today considered obsolete. It once was the low-noise choice in the 70s (40 years ago), but not any more - and if you can find them they're often quite expensive. The BC109 (A, B, or C) has a noise figure of 4.0db typically, whereas a BC549/BC550 (again A, B, or C) has a typical noise figure of 1.2db. Forget the BC109.

The A/B/C suffix only denominates the selected gain range, it has nothing to do with noise.

Depending on the signal source impedance a FET might make a better choice, something like a 2SK30, noise figure 0.5db and dirt cheap.

Steve A.

Re: BC109C Replacement?

PostPosted: Wed May 20, 2015 9:56 pm
by Lawnboy
Thanks guys. I actually confused myself a bit, the original design for the head amp called for a BC172 (see Vol 38 No 1 p 13), but I didn't have any so I had asked Jeremy if I could use a '109 instead. But even the 172s are hard to find! The transistor follows an op ampĀ  (I am using a TL072 instead of the CA3130) so I guess it's there as a second stage of gain? Would a FET have any advantage here?

Re: BC109C Replacement?

PostPosted: Wed May 20, 2015 11:18 pm
by Steve Anderson
None whatsoever. The noise is going to be dictated by the op-amp not this transistor. You could use any common or garden NPN here, but the biasing will need a tweak (it's horrible, unreliable, device dependent and temperature sensitive). Dump the 741 and use something more modern - the other half of the TL072 you suggested...makes sense to me! If you wish use a single TL071 and simply an emitter follower after the tranny instead of the op-amp.

Steve A.

Re: BC109C Replacement?

PostPosted: Thu May 21, 2015 1:35 am
by Lawnboy
Steve, interesting you should mention dropping the 741. I asked Jeremy about using the other half of the TL072, and he said that that was the original idea, but upon test there was a significant reduction in speed, which he reckoned was due to capacitive coupling within the IC. For my prototype I am using a TL072 and a LM358. One circuit will use 4x BPW34s behind a "4-up" LED lens and the other will have Silonex cell paired with a reflector. As you can guess this is for a flying spot pickup.

Re: BC109C Replacement?

PostPosted: Thu May 21, 2015 11:36 am
by Steve Anderson
Lawnboy wrote:...but upon test there was a significant reduction in speed, which he reckoned was due to capacitive coupling within the IC.

I would very much doubt that, without looking it up the channel isolation figure in these sort of op-amps is in the order of 120db, a million to one. There must have been something else involved here.

Steve A.

Re: BC109C Replacement?

PostPosted: Thu Apr 28, 2016 6:51 am
by Lawnboy
Hi guys. Last weekend I tried building an experimental version of Jeremy's amplifier using both sides of a TL072, and unfortunately got the same result that Jeremy reported in the newsletter- a blurring along the line caused by slow speed response. Any ideas as to what would cause that? I also swapped the BC109C with an ordinary 2N2222. No other components were changed.

My goal is to get a high quality, low complexity head amp. Steve, you mentioned dropping the second op amp and using an emitter follower. Would that be the best way to go?

Re: BC109C Replacement?

PostPosted: Thu Apr 28, 2016 5:43 pm
by Steve Anderson
Unfortunately which way round to connect the photo-diode has been omitted from the article - unless there was an 'errata' published later. If you end up with negatives it's obviously wrong.

There should be no reason for bandwidth problems from any standard (non-bipolar, i.e a J-FET or CMOS) op-amp used here. Where the issue may be is around the input to the first stage with the 4Meg feedback resistor. Reduce any stray capacitance as much as possible. No screened cables and the first stage should be within a few cm of the detector/photodiode. Millimetres would be better.

The TL072 has a gain-bandwidth product of a minimum of 2.5MHz. So the first stage can have a gain (in current) of 250 before running out of steam for NBTV applications. Just 4pF will cut any response to 10kHz with a 4Meg feedback resistor. And 4pf is a very small amount of capacitance. Probably equal to the op-amp's input capacitance alone, let alone any other board or wiring strays.

1/(2pi*10000*4,000,000) = -3db @ 10kHz...or 4pF

Steve A.

I might add that although Texas Instruments tout the TL071/72/74 series as being 'low-noise', they're not. They have the same noise spec as the TL081 etc. series.

If you really want low-noise go for a bipolar LM394 (virtually unobtainable today) or even better a SSM2210. (Analogue Devices - be prepared to open your wallet wide).

Though with tongue-in-cheek the tube-heads might suggest an EF86...which isn't as daft as it sounds.

Re: BC109C Replacement?

PostPosted: Thu Apr 28, 2016 8:27 pm
by Harry Dalek
When i was testing what i had handy the 5534 came out tops in my junk box but the LT1115 sounds like worth testing once i get my hands on one .

Re: BC109C Replacement?

PostPosted: Fri Apr 29, 2016 12:50 am
by Lawnboy
Hi Steve, I'm really not sure how much of a role capacitance is playing in this circuit. The other photocell I built uses a TL072/BC109C/LM358 combo with a Silonex cell and I get impressive results from that, and the quoted junction capacitance for that cell is 2nf, not to mention the 3cm+ leads.

I will double check the cell polarity when I get a chance later today.
EDIT: The positive terminal connects to pin 5 (the non-inverting input) and the negative to pin 6.

Re: BC109C Replacement?

PostPosted: Fri Apr 29, 2016 7:52 pm
by Steve Anderson
Well, 2nF of junction capacitance certainly dominates. Any applied reverse bias?

The NE5534 and the LM833 are very similar in noise specs, not really much to choose between them. But of course they are complete bipolar op-amps whereas the LM394 and the SM2210 are two low-noise transistors (only) in a package that looks for all the world like a 741. I still have a bunch of LM394s in the hermetic sealed TO99 package. Where I got them from I can't recall. I expect they were horrendously expensive though I assure you I didn't pay the full price for them! Cricklewood Electronics (UK) have them listed at 18 Pounds Sterling (+VAT of course).

Steve A.

I've got a feeling that the LM394s I have were from a low-noise sound pre-amp used in a movie projector that were never completed. Maybe the company went bust. (I've had them decades). There are still very few devices that can beat them on noise performance. One of the nearest contenders is a J-FET, the 2SK30 (Toshiba), a single device though and dirt cheap.

Re: BC109C Replacement?

PostPosted: Sat Apr 30, 2016 1:21 pm
by Lawnboy
The cell is wired in just as the diagram shows. I am not smart enough to mess with the biasing yet! :D Don't forget that this is a virtual earth config, so the cells are practically running short circuit.

At first I wasn't overly concerned with noise, since this preamp was for a flying spot scanner and would have plenty of signal available compared to being hid behind a Nipkow disc. However if I want to light the scene beyond a few feet, I suppose I should have minimal noise. Without going for a complete redesign of the preamp, would a NE5532 be a drop-in replacement to a TL072?

What would be the best way to combine the signals from two separate preamps in the same container? It just occurred to me that wiring them in parallel (which is how it is hooked up in one of the prototypes) is probably not the best approach. I have been using dual chips mostly because I already have them, and why not include two sensors for the difficulty of one? :D