Resistor range adjustment circuit

T

Tim Angus

Guest
I better start by apologising in advance. My electronics expertise doesn't
extend much beyond what I learnt at high school, and that was 6 years ago.
As such I may be found to be talking out my ass to a certain extent.

I doubt the background to my problem is particularly important, but I
shall state it regardless. I am currently attempting to build input
devices for a Roland electronic drum module. One of these input devices
controls the hi hat. The commercial hi hat pedal senses foot pressure
using a force sensing resistor. I've searched high and low for a place to
get such devices in the UK in low volume, but I've not come up with
anything.

I did come up with an alternative however, in the form of part of the PCB
from an XBOX games console controller. It has a set of buttons which
control analog devices behaving relatively similarly to what the afore
mentioned Roland drum module expects on its pedal input. But not similarly
enough.

This input measures the resistance across its 5V line and ground. The
device which you connect to the input must have a resistance range from at
least 50KOhms to 0Ohms (or very close). The device I've hacked up from the
games controller has a range of infinity Ohms (i.e. open) to about 5KOhms.
As a result I need to transform its resistance to map to a range which the
drum module understands.

I've come up with the following naive circuit:

5V ----------------------
| |
F |
S |
R |
| /
|-------|
| \.
T |
V |
R |
| |
0V ----------------------

Where FSR is the hacked up force sensing resistor, TVR is a trimming
variable resistor, and the jumble on the right is an NPN transistor. So,
this sort of works as far as the range of the resistance is correct. When
pressing the FSR though, as soon as the base voltage passes 0.6V the
transistor closes rapidly. In other words the sensitivity of the device is
lost so much so that it is barely better than the momentary switch that it
replaces in my current setup.

So basically, am I doing anything right? Is there some adjustment to this
circuit, or entirely different circuit that would satisfy my requirement
to transpose the resistive range yet retain the sensitivity of the device?

Thanks, Tim.
 
Tim Angus wrote:


This input measures the resistance across its 5V line and ground. The
device which you connect to the input must have a resistance range from at
least 50KOhms to 0Ohms (or very close).
Do you know if the Roland electronic drum module passes only DC through
the external control resistor? Or does it actually apply an audio signal
across the external resistor (kinda like a volume control/fader where
the audio signal actually appears across the resistor)?

The solution to matching your Xbox resistors to what the Roland expects
depends on the answer to the questions above.

MikeM
 
On Mon, 15 Mar 2004 16:20:00 -0700, MikeM wrote:
Do you know if the Roland electronic drum module passes only DC through
the external control resistor? Or does it actually apply an audio signal
across the external resistor (kinda like a volume control/fader where
the audio signal actually appears across the resistor)?
It's DC only at 5V.
 
Tim Angus wrote:

On Mon, 15 Mar 2004 16:20:00 -0700, MikeM wrote:

Do you know if the Roland electronic drum module passes only DC through
the external control resistor? Or does it actually apply an audio signal
across the external resistor (kinda like a volume control/fader where
the audio signal actually appears across the resistor)?


It's DC only at 5V.
How about turning your proposed transistor circuit upside down.
Use a PNP transistor instead of the NPN.
Put the FSR between the base of the transistor and ground.
Put the TVR between the base and +5V.
Ground the collector of the PNP.
Connect the emitter of the PNP to the Roland.

This configuration is called an emitter follower. It
has no voltage gain, just power-gain, where a small
base current controls a larger current flow in the
emitter. In other words, it converts the high resistance
of the FSR to a much lower effective resistance at
the emitter.

MikeM
 
On Mon, 15 Mar 2004 16:49:31 -0700, MikeM wrote:
How about turning your proposed transistor circuit upside down.
Use a PNP transistor instead of the NPN.
Put the FSR between the base of the transistor and ground.
Put the TVR between the base and +5V.
Ground the collector of the PNP.
Connect the emitter of the PNP to the Roland.

This configuration is called an emitter follower. It
has no voltage gain, just power-gain, where a small
base current controls a larger current flow in the
emitter. In other words, it converts the high resistance
of the FSR to a much lower effective resistance at
the emitter.
I've tried using a PNP as you suggest and while the sensitivity of the FSR
is retained, the range is little better than just using the FSR connected
directly to the Roland unit. I guess because the FSR never goes below
5KOhms the voltage divider will never produce a low enough voltage to
completely close the transistor. It's very important that when the FSR
hits its lowest resistance, the resistance of the overall circuit is as
close to 0 as possible.

Could a FET be the answer to my problems? I've read up a little, but I'm
still a little lost. Am I right in saying the resistance of a (J?)FET is
inversely proportional to the voltage on its gate pin. If so is it a
roughly linear relation?
 
In article <pan.2004.03.16.04.23.31.67436@NOSPAMsms.ed.ac.uk>, Tim Angus
<T.M.Angus@NOSPAMsms.ed.ac.uk> writes
On Mon, 15 Mar 2004 16:49:31 -0700, MikeM wrote:
How about turning your proposed transistor circuit upside down.
Use a PNP transistor instead of the NPN.
Put the FSR between the base of the transistor and ground.
Put the TVR between the base and +5V.
Ground the collector of the PNP.
Connect the emitter of the PNP to the Roland.

This configuration is called an emitter follower. It
has no voltage gain, just power-gain, where a small
base current controls a larger current flow in the
emitter. In other words, it converts the high resistance
of the FSR to a much lower effective resistance at
the emitter.

I've tried using a PNP as you suggest and while the sensitivity of the FSR
is retained, the range is little better than just using the FSR connected
directly to the Roland unit. I guess because the FSR never goes below
5KOhms the voltage divider will never produce a low enough voltage to
completely close the transistor. It's very important that when the FSR
hits its lowest resistance, the resistance of the overall circuit is as
close to 0 as possible.

Could a FET be the answer to my problems? I've read up a little, but I'm
still a little lost. Am I right in saying the resistance of a (J?)FET is
inversely proportional to the voltage on its gate pin. If so is it a
roughly linear relation?
No...

just to veer off in another direction, this device you've got out of the
XBOX, what is it exactly? has it got any other terminals?

If you put a 50K resistor across the XBOX device, you'd get 50K down to
4.5K, which is about 90% of the range you need, would that work?
--
Tim Mitchell
 
Tim Angus wrote:


I've tried using a PNP as you suggest and while the sensitivity of the FSR
is retained, the range is little better than just using the FSR connected
directly to the Roland unit. I guess because the FSR never goes below
5KOhms the voltage divider will never produce a low enough voltage to
completely close the transistor. It's very important that when the FSR
hits its lowest resistance, the resistance of the overall circuit is as
close to 0 as possible.
Ok, what voltages are available? Any negative voltages?
 
On Tue, 16 Mar 2004 16:09:23 +0000, Tim Mitchell wrote:
It's very important that when the FSR
hits its lowest resistance, the resistance of the overall circuit is as
close to 0 as possible.

just to veer off in another direction, this device you've got out of the
XBOX, what is it exactly? has it got any other terminals?
It's a section of the PCB, with interlocking finger type conductors, which
are rounded on the surface. On top of that is a resistive material
(presumably doped in some way). When pressure is applied to the top the
resistive material bends round the rounded conductors thus shorting them.
The more pressure is applied the shorter the physical distance between the
conductors and hence the lower the resistance.

If you put a 50K resistor across the XBOX device, you'd get 50K down to
4.5K, which is about 90% of the range you need, would that work?
No, due to the quote above. The open end of the resistance range is pretty
unimportant -- it could be an open circuit or it could be as low as
50 KOhms. Much more important is the lower end approaching 0 Ohms
 
On Tue, 16 Mar 2004 09:35:07 -0700, MikeM wrote:
I've tried using a PNP as you suggest and while the sensitivity of the FSR
is retained, the range is little better than just using the FSR connected
directly to the Roland unit. I guess because the FSR never goes below
5KOhms the voltage divider will never produce a low enough voltage to
completely close the transistor. It's very important that when the FSR
hits its lowest resistance, the resistance of the overall circuit is as
close to 0 as possible.

Ok, what voltages are available? Any negative voltages?
Only the +5V (presumably used for logic inside the module) and 0V. I could
conceivably use another power source, but I'd rather avoid that if
possible.
 
Tim Angus wrote:

Only the +5V (presumably used for logic inside the module) and 0V. I could
conceivably use another power source, but I'd rather avoid that if
possible.
Ok, here is the best I can do without a better definition of the problem

I have cobbled together something which you can play with to get a final
circuit.

1. Download and install LTSpice/SwitcherCADIII from http://www.linear.com/software/

2. cut and paste everything below the _____________ into a text file
called "res.asc".

3. Run SwitcherCADIII

4. File/OPEN res.asc

5. click on the little running man

6. After it runs, click on waveform at V(vin)

This shows how the voltage at the Rolands input pin (Vin, y axis) varies as
the Xbox resistance (x axis) changes from 4.5K to 50K, for different
positions of the 500K pot (the different color traces).

Unfortunately, the transfer curve is effected by the threshold voltage of
the FET, and to a lesser extent by whats inside the Roland (R1), so I
cant give you definite values for the Pot position (pot), and R4. However,
if you figure out how to change the simulation to reflect your conditions,
you can resim it at your leasure to get the transfer curve you want...

Mike Mladejovsky

________________________________________________________________________________

Version 4
SHEET 1 1048 680
WIRE -16 256 80 256
WIRE 240 272 240 304
WIRE 160 256 192 256
WIRE 240 176 240 64
WIRE -256 -96 624 -96
WIRE 720 32 720 64
WIRE -256 256 -208 256
WIRE 192 64 240 64
WIRE 240 64 240 0
WIRE 112 64 80 64
WIRE -128 256 -96 256
WIRE -48 64 -80 64
WIRE 80 256 160 256
WIRE 32 64 80 64
WIRE 80 64 80 256
WIRE -256 -96 -256 256
WIRE 240 64 288 64
WIRE 720 -48 720 -96
WIRE 624 0 624 -96
WIRE 624 -96 720 -96
WIRE 496 0 240 0
WIRE 576 0 624 0
WIRE -80 64 -80 96
FLAG 240 304 0
FLAG 720 64 0
FLAG 160 256 g
FLAG -80 96 0
FLAG 288 64 Vin
SYMBOL nmos 192 176 R0
SYMATTR InstName M1
SYMATTR Value NDS9410A
SYMBOL res 592 -16 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 32 56 VTop 0
SYMATTR InstName R1
SYMATTR Value 1000
SYMBOL Misc\\battery 720 -64 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V1
SYMATTR Value 5
SYMBOL res 0 240 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 32 56 VTop 0
SYMATTR InstName R3
SYMATTR Value {Xbox}
SYMBOL res 48 48 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 -62 62 VTop 0
SYMATTR InstName R2
SYMATTR Value {500000*pot}
SYMBOL res -112 240 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 32 56 VTop 0
SYMATTR InstName R4
SYMATTR Value 54.9K
SYMATTR SpiceLine tol=1 pwr=0.1
SYMBOL res 208 48 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 -60 58 VTop 0
SYMATTR InstName R5
SYMATTR Value {500000*(1-pot)}
TEXT 328 312 Left 0 !.STEP param Xbox 4.5K 50K 2.5K param pot 0.15 0.3 0.01
TEXT 320 352 Left 0 !.op
TEXT 368 120 Left 0 ;Notes: R1 is a guess of what is inside the Roland\nV1 is
the 5V logic supply\nR2/R5 is a 500K pot, sets the range of the transfer
curve\nR3 is the Xbox variable resistor, 4.5K- 50K\nR4 sets the slope of the
transfer curve\nThe threshold voltage of M1 effects the slope/range
 
On Tue, 16 Mar 2004 12:41:20 -0700, MikeM wrote:
Ok, here is the best I can do without a better definition of the problem
snip
Unfortunately, the transfer curve is effected by the threshold voltage of
the FET, and to a lesser extent by whats inside the Roland (R1), so I
cant give you definite values for the Pot position (pot), and R4. However,
if you figure out how to change the simulation to reflect your conditions,
you can resim it at your leasure to get the transfer curve you want...
Thanks very much for your help, I'll pop down to Maplin tomorrow and pick
up a few FETs to tinker with. I hadn't come across LTSpice/SwitcherCADIII
before, it's a nice bit of kit. I've used oregano on Linux from time to
time, but this seems somewhat superior.

This whole adventure has woken me up to just how little I know about
analogue electronics. My digital electronics is alright thanks to some low
level computer design courses I took during my CS degree. Supposing I
wanted to teach myself analogue electronics from nearly the ground up, has
anyone any recommendations for books or websites I should be looking at?
 
"Tim Angus" <T.M.Angus@NOSPAMsms.ed.ac.uk> wrote in message
news:pan.2004.03.18.00.00.30.944404@NOSPAMsms.ed.ac.uk...
This whole adventure has woken me up to just how little I know about
analogue electronics. My digital electronics is alright thanks to some low
level computer design courses I took during my CS degree. Supposing I
wanted to teach myself analogue electronics from nearly the ground up, has
anyone any recommendations for books or websites I should be looking at?
Art of Electronics, by Horowitz and Hill.

Also, given your music-related interests, any of the electronics project
books by Craig Anderton.
 
Tim Angus wrote:

This whole adventure has woken me up to just how little I know about
analogue electronics. My digital electronics is alright thanks to some low
level computer design courses I took during my CS degree. Supposing I
wanted to teach myself analogue electronics from nearly the ground up, has
anyone any recommendations for books or websites I should be looking at?
Art of Electronics, Horowitz and Hill, ISBN 0-521-37095-7

Win is a frequent contributor here ...

Mike Mladejovsky, PhD EE
 

Welcome to EDABoard.com

Sponsor

Back
Top