Geofex humfree splitter, offset voltage with lm833n

[skumberg]

Hi,

I have built the humfree splitter according to the latest schematic with just some minor differences. http://www.geofex.com/FX_images/TransformerSplitter.pdf

The differences are:
-I used a capacitor on the input in series with the 10k.
-30k resistor in series with 2.2uf to reduce gain. Might adjust this later on.
-no switching

With the LM833n I get a 0.4V dc offset on the direct output and + input. I assume this is from bias current through the 2M2. A TL072 has no offset voltage.

Should I use a DC blocking capacitor instead of the 1k in series with the direct output? Or can I use the TL072?

Either way the pdf might need an update. I'm very grateful of the projects R.G. publish and just want to point out a minor adjustment.

R.G. by the way I have the SCD preamp done in a stompbox. It was over a year we spoke about it. Maybe I should post my results in a thread here later on, or I'll pm the resistor changes I did in the schematic?

[R.G.]

With the LM833n I get a 0.4V dc offset on the direct output and + input. I assume this is from bias current through the 2M2. A TL072 has no offset voltage.

Should I use a DC blocking capacitor instead of the 1k in series with the direct output? Or can I use the TL072? [/quote]
Bizarre. The datasheet does say the bias current could be big enough to do that, OK, but in practice I've never had one go that far. I would drop the 2.2M down to 1M. If that drops the offset to 0.4V, it's a bias current issue.

In that case, you can use either approach. I like the 833 a lot for its abilities to drive larger currents into loads, but if the TLO72 does what you want, use it. Also, you could as you say put a cap in series with the 1K to the direct output. The offset isn't really an issue except on the direct output.

R.G. by the way I have the SCD preamp done in a stompbox. It was over a year we spoke about it. Maybe I should post my results in a thread here later on, or I'll pm the resistor changes I did in the schematic?

Whichever you like. I found my populated SCD preamp board in a bench cleanup a couple days ago and **someday** I'll get time to try it myself.  Maybe PM me if I did something really dumb...   

[skumberg]

Thank you for the suggestions. The offset went down to about 0.2V when I used a 1M on the input. So I added a 10uf output capacitor. However this LM833N from ST electronics is giving me more problems. There is a whistling that I can change the frequency of by turning the guitar controls and by waving my hands close to the strings. This is not present on the TL072IP I'm using which is also from ST. It's not very loud but very noticeable.

Edit: Whistling problem solved with a 220pf Cf capacitor across the 11k feedback resistor.

The original SCD schematics we had was very blurry, searching the net for pictures have given me the correct values. And thankfully we now have additional help from this site. The values agree with what I have found. And now we have all the versions of the MRB circuit I think. I also built the mixer circuit since tha gain was to low without it.
http://www.backfromthesixties.co.uk/siteimages/vox%20supreme%20mk1%20preamp%20schematic.jpg
http://www.backfromthesixties.co.uk/siteimages/Vox%20Supreme%20MkIII%20Preamp%20Schematic.jpg

[gritz]

We had a thread a week or two back asking if equivalent opamps from different manufacturers could behave differently. There wasn't a definitive answer, but depending on manufacturer some LM833s (those from Texas Instruments and National Semiconductor spring to mind) have a specified minimum supply voltage of ±5V and while ±4V5 is only 10% less, it's still possible imo that some LM833s might get edgy - especially into a reactive load (as we have here). That said, the Nat. Semi. ones I have here have always behaved ok at 9V - although I don't really beat on them.

The offset voltage thing is fairly typical of what I've experienced here with bipolar input opamps connected to high value input resistors - and it can be trimmed out if it's a problem. Perhaps if you're not using the other half of the opamp then you could rejig and have half a TL072 driving each transformer instead. [/politician's kludge mode]

Edit: I guess it rather depends what you're going to plug into the input of the splitter. If it's a passive guitar and nothing else then you don't really need a coupling cap at the input. If it's f.x. then then you'll want a coupling capacitor, but you can afford to go down on the input impedance a little and the offset will disappear - although there's always the possibility that someone else will plug in *something unexpected*...  

[skumberg]

Hi gritz thanks for your thoughts. I've been playing with it and I think it sound alright at the moment. There will be pedals at the input so I want the input cap, but see your point, a lower source resistance would reduce the bias and offset voltage. I might get another brand of lm833 in the future to try out.

I'm using an icl7660s to produce +9V and -9V for this pedal. It has the frequency doubling function and higher input voltage of 12V(max 13V but not recommended). So the 0.4V offset should not matter much in terms of headroom. I just hope there isn't any other problems I need to look out for. As I understand it the LM833 should be able to handle the load easily, two transformers and one direct output into 3 amplifiers.

I used a BB box with a small teco steel box inside to act as an extra shield. The circuit board with transformers are all mounted inside the teco with cut outs for the cables. I also used the PNP switching methode in one of RGs latest pdfs.

[gritz]

Hi skumberg: ah, you're using a split supply and voltage doubler, so notwithstanding any possible issues with filtering / noise pickup / layout and whatnot (which you sound like you're getting on top of) then the LM has plenty of volts to play with. Let us know how it goes mate.

Edit: because the LM833 is capable of driving low impedances I think it's worth making sure that there is sufficient power supply decoupling as close as possible to the chip.

[skumberg]

Well it's a tight layout with a 55*50mm board(17*19 perf holes) but I think I have done a good job seperating the power and signal ground.

The switch I'm using is this http://www.geofex.com/FX_images/PNP_power_switching.pdf. The values are Rbase 5K1 and Rjtbs 100K which I think will give enough current and low enough saturation voltage. After the PNP switch I have a 47R and 100uf filter plus a 12V zener to protect the ICL7660s. Other than that I just use the two 10uf specified in the icl7660s sheet to get + and - supply. Sometimes I see people putting both electrolytic and ceramic caps to ground from the + and - supply. But these are not shown in the datasheet applications.