FET input stage for instruments

[rockinrob86]

Looking at the ASO98 schematic, at the end of the part in the square, where it says +12db , + - to xformer, how do I calculate what the impedance is?

I'm just guessing that it has been converted in this circuit to low impedance (10K?) and at that point, the two circuits would be feeding the blend knob (which would be 10k to match?), and then from there hit the XLR cable to go to the floor. 

At the floor, I would need another gain stage of 12-24db,  low impedance impedance input, and output 600 ohms into my -12db insertion loss EQ circuit, before feeding a final makeup stage and a transformer to convert to balanced and go to a mixing board mic pre in.

Does this sound right?

 

[clintrubber]

Looking at the ASO98 schematic, at the end of the part in the square, where it says +12db , + - to xformer, how do I calculate what the impedance is?

I'm just guessing that it has been converted in this circuit to low impedance (10K?)

It's much lower. The effective output 'impedance' is the series combination of the Q1,Q2 part plus R5 (100).
R7 is in parallel to that, so lowers it slightly an additional bit.

Being lazy for now to think about the output-Z of the Sziklai-section (Q1, Q2 & feedback), let's say that for ease of thinking it's near zero  (see the other file as004.pdf: you want to have the source impedances feeding the balanced cable over there to be equal for maximum rejection of disturbances, and they've taken R5 = R4 over there).

OK, so the output impedance say 100 Ohm-ish, mainly determined by of  R5.

and at that point, the two circuits would be feeding the blend knob (which would be 10k to match?), and then from there hit the XLR cable to go to the floor. 

Matching impedances isn't relevant here. You have two 100 Ohm signal-sources (each PU an own buffer), blend these with a pot of say 5k, or 10k (LIN). If you only have a 2k5 that would work as well.

At the floor, I would need another gain stage of 12-24db,  low impedance impedance input, and output 600 ohms into my -12db insertion loss EQ circuit, before feeding a final makeup stage and a transformer to convert to balanced and go to a mixing board mic pre in.

Does this sound right?

Several ways to do this.

First of all, the input impedance at the floor, which loads the wiper of the blend pot  could be as low as say 47k, but in case you want to use it for directly plugging in passive pickups you might want to use 470k or 1M.

You could indeed go balanced with a TX, or use an inverting opamp stage, or simply use the same trick as as004.pdf does: mimicing a no-signal source by C3 + R4.
Remember that the benefits of balanced for disturbance rejection are due to identical source resistances, not by having live signals on each 'signal'-conductor. One of them can be 'silent'.

before feeding a final makeup stage and a transformer to convert to balanced and go to a mixing board mic pre in.

Not sure you need al this gain to feed a mic pre in. No idea how much signal your PU's generate, but if all this gain is indeed needed it might be better to shift that required gain more upstream in the signal chain.

Regards

 

[rockinrob86]

This is my first attempt at pcbs design...

How does this look?  https://easyeda.com/rosenton/AS098_Stereo_Preamp-f02f308980574d7b8cea1ff680c06ea6

I copied the jensen schematic into the system, and then converted it to a pcb.  I am not really sure how the layout is.  Also, I didn't mark everything on the schematic, but it follows the jensen one exactly, other than the input, output and +9V are on connectors.  I'm planning on wiring this pcbs into a small box by the guitar, with the output feeding a 10K blend pot, and then sending a mono out from the small box located at the guitar.  May also add a volume pot.  This would probably go into an XLR cable to a box on the ground with the power supply and maybe some other goodies, but first thing first I need this buffer/preamp figured out!

Does anyone see any rookie mistakes on the schematic/pcb?

 

[clintrubber]

Hi,

Schematic looks good wrt the Jensen-schematic.  One error: the node which should be common to  Q4, R10, C5
isn't currently so.

The use of 'x' to indicate no connection of crossing wires is non-conventional, but clear what it indicates.

For the layout one could suggest to do a compacter layout for one 'channel' and then simply duplicate (copy) that next to it.
Because this is low frequency stuff, it'll be less critical than more HF stuff , but still.

I hadn't heard of easyeda before, thanks for bringing that up. 

Had a short look at what in the end a few fabricated PCBs including shipping to Europe would cost, but it wasn't that clear to me. Or I simply couldn't believe the $2 price for each, even at low numbers. 

Have a good weekend

 

[metalb00b00]

I hadn't heard of easyeda before, thanks for bringing that up. 

Had a short look at what in the end a few fabricated PCBs including shipping to Europe would cost, but it wasn't that clear to me. Or I simply couldn't believe the $2 price for each, even at low numbers. 

Have a good weekend

easyEDA boards are more or less the same quality as PCBWay.
I just had them manufactured 3 of my PCB designs last week.

1st PCB design, 10 pcs, 30 mm x 30 mm, cost $2.
2nd PCB design, 10 pcs, 80 mm x 30, mm cost $5. Didn't even change the solder mask color, or anything else.
3rd PCB design, 5 pcs, 360 mm x 60 mm cost $16 (PCBWay cost $40)

2 working days to make, and another 2 days with DHL for them to arrive at my front door.

 

[clintrubber]

easyEDA boards are more or less the same quality as PCBWay.
I just had them manufactured 3 of my PCB designs last week.

1st PCB design, 10 pcs, 30 mm x 30 mm, cost $2.
2nd PCB design, 10 pcs, 80 mm x 30, mm cost $5. Didn't even change the solder mask color, or anything else.
3rd PCB design, 5 pcs, 360 mm x 60 mm cost $16 (PCBWay cost $40)

2 working days to make, and another 2 days with DHL for them to arrive at my front door.

No recent experiences here wrt pricing, but that sounds more than reasonable, not ? Thanks for sharing!

 

[clintrubber]

[quote author=rockinrob86]
I bought a stereo PUTW #54 SBT (piezo film strips that glue to the bridgeplate, sound board transducers).  These have very low output[/quote]

Not familiar with these pickups,
but just checking that these generate their signal without any additional requirements for some DC-biasing etc ?

Otherwise that better be added to the PCB right away.

Bye

 

[rockinrob86]

Unfortunately, it looks like I screwed it up. 

There is no continuity between any of the traces on my PCBs.  It looks like I missed that step in the PCB design software?  I'll have to dig into it this weekend and figure it out, any tips? 

Was I supposed to do that as a different layer?  Right now the PCBs have a smooth appearance on the top and bottom, with no sign of traces.  I assumed they were just on the inside when I first saw them, but my DMM confirmed, no continuity

 

[clintrubber]

Unfortunately, it looks like I screwed it up. 

There is no continuity between any of the traces on my PCBs.  It looks like I missed that step in the PCB design software?  I'll have to dig into it this weekend and figure it out, any tips? 

Was I supposed to do that as a different layer?  Right now the PCBs have a smooth appearance on the top and bottom, with no sign of traces.  I assumed they were just on the inside when I first saw them, but my DMM confirmed, no continuity

Might be best to post some pics.

For now: the layout-viewer for your PCB-design seemed to indicate you went for a two-layer board?
You didn't see connectivity between front & back possibly ?
No problem there, some PCBs use through hole connectivity (there's a better word for this...), but yours probably didn't.
Still no problem, as soon as a component lead is inserted AND soldered at both sides (!) the connection is made.

But again, perhaps not what you meant - so best to post some pics. 

 

[rockinrob86]

I thought that was how it was, but I checked it with my meter.  None of the ground points are connected to each other.  Following through the schematic, it looks like not a single point that should have connectivity does.  I think I made a nice layout on the PCB, with a pretty picture of the component on the top, and then a nicely plated hole, and then nothing else.

I'm just going to have to play with the software some and see - I didn't add any layers to it on the software, and I made sure the little blue lines were connected to the right points.  I'm thinking maybe I needed to do that part in a different layer or something?

 

[clintrubber]

Hi,

To be of any reasonable help we could use some pics, otherwise we're just guessing. Understanding you have fabricated PCBs in front of you ?

 

[rockinrob86]

I don't have them in front of me, I can take pictures later.  However, it just looks like a normal PCB with no sign of traces on the bottom.

The design comes from here, as listed above:
  https://easyeda.com/rosenton/AS098_Stereo_Preamp-f02f308980574d7b8cea1ff680c06ea6

I made the connecting lines on the pcb but looking at it now it looks like maybe I should've copied over those lines while selecting bottom layer, or at least this is my hypothesis

 

[clintrubber]

However, it just looks like a normal PCB with no sign of traces on the bottom.

Sorry, not trying to be funny here, but that's then not a normal PCB. Did the fab do some visual inspection before they shipped them to you?

OK, let's have another look at this when you've had the opportunity to post pics.

Bye

 

[PRR]

{sigh} You drew "ratlines", conceptual connections, net-lists.

What you did not draw is "Tracks", actual strips of copper.

See attached. After a moment of staring at the editor, I knew the skinny lines were concepts, but saw the tool top-left of the toolbox which does look like 1970s tape&donuts masking. And poking around, it did the 90 and 45 degrees typical of basic trace layout. I drew a big "?" over your plan. I knew it was doing something, because a couple times it popped-up the question "Are you sure you want to merge Q2 and R5?", meaning it was really making a connection, and not one shown in net-list or ratlines.

 

[warpie]

If my eyeball is correct, Q2 will wiggle up and down to try and maintain the drain voltage at near DC, which sets a constant current through R5. If R5 = 1K, and the DC bias current is set to 0.7mA or below, doesn't this force Q1 off always?  It's also not clear to me how Q1 is helping (even assuming the bias works out to turn it on) - even if gain is 1, the gate would need to be driven with one hell of a voltage swing to place the output up near a +24V rail.

This topology makes more sense to me, as the output current wouldn't be limited by what Q2 can sink (aka a lower impedance load):

Thanks for this schematic. Just a few basic questions if you don't mind

1) Would that DI fit in front of a mic pre transformer? I'm building my first neve clone and I'm looking for a suitable DI circuit to include in the same 1u rack case. BTW, have you used this circuit? If so, does it sound any good (on passive and active gtrs/basses/etc..)?

2) What is the impedance (2M2?)

3) Is Q3 optional?

Thanks again!

 

[PRR]

> What is the impedance (2M2?)

"Infinite".

C2 bootstraps R3 from the nearly unit gain point at Q1 Source. Say Q1 runs at gain of 0.9; then R3 appears 10 times (1/(1-0.9)) larger, 20Meg. Q1's gain may be higher. IAC, we will connect with a cable, which has capacitance. 3 feet of cable is 100pFd, which is an impedance lower than 20Meg for anything over 83Hz; ~~200K impedance at 8KHz. So whether Zin is really 2Meg or 20Meg hardly matters.

 

[warpie]

Thanks for the explanation PRR!

Do you know in which case Q3 is needed?  As I mentioned, I'd like to to add this circuit in front of a mic preamp (as a Hi-Z input) so the "long cable" situation won't apply here I guess.

 

[abbey road d enfer]

Do you know in which case Q3 is needed?  As I mentioned, I'd like to to add this circuit in front of a mic preamp (as a Hi-Z input) so the "long cable" situation won't apply here I guess.

The output impedance of the source-follower is roughly 1/gm, so about 1-2 kOhms, which is not really fine for driving a mic input, particularly if it has a transformer front-end. Adding Q3 reduces the impedance to a few dozen ohms, which is much better in terms of frequency response and noise. However, it may be necessary to adjust R6 in order to optimize headroom.

 

[Jarno]

I may have a board for this circuit, which works. Let me know if you are interested and I'll check my stash of PCB's.

LOL, in the meantime, I did learn how to do a ground pour  ;D

 

[warpie]

The output impedance of the source-follower is roughly 1/gm, so about 1-2 kOhms, which is not really fine for driving a mic input, particularly if it has a transformer front-end. Adding Q3 reduces the impedance to a few dozen ohms, which is much better in terms of frequency response and noise. However, it may be necessary to adjust R6 in order to optimize headroom.

Thanks, that makes some more sense to me Is it too much to ask how do I go about calculating the output impedance (i.e how did you calculate the "few dozen ohms")? I would assume that the lower the R6 the higher the headroom but how do I choose the correct value.

I may have a board for this circuit, which works. Let me know if you are interested and I'll check my stash of PCB's.
LOL, in the meantime, I did learn how to do a ground pour  ;D

Are you referring to me or to the OP?

BTW, I haven't tested it but what I like about it is that it has a very high input impedance so i guess it's appropriate for both passive and active guitars?