Slope limiter to slope "pass" conversion

[xmat]

Hi, attached u shall find a circuit that will limit the maximum rate of change of an input signal, NOT the frequency it passes..(see also graph)

Considering the limitations of the circuit (i.e R1<<R2) and that the "slope limit" is set by R2C1,

I want to build a similar circuit that does the exactly opposite job:

It allows rapidly-varying (parts of) signals to pass unaffected, while "filtering" slow-varying (parts of) signals...

Any hints/schematics to do that??

Thanx in advance,

xmat

 

[Roff]

See below.

 

[Roff]

Here is a way that does no lowpass filtering on the input. It is a little more complex.

 

[Optikon]

Here is a way that does no lowpass filtering on the input. It is a little more complex.

I've attempted to do precision slew control with a similar circuit in the past. It used the same current sources through diodes to a capacitor idea. I couldnt reconcile the fact that the diodes would not equally share the current (and hence error in slew control) this certainly fell apart over temperature. Balancing the voltage on both sides didnt entirely eliminate the issue either. Other than matched diode pairs, do you have an idea for making this precision (like .01% control)???

 

[Roff]

Here is a way that does no lowpass filtering on the input. It is a little more complex.

I've attempted to do precision slew control with a similar circuit in the past. It used the same current sources through diodes to a capacitor idea. I couldnt reconcile the fact that the diodes would not equally share the current (and hence error in slew control) this certainly fell apart over temperature. Balancing the voltage on both sides didnt entirely eliminate the issue either. Other than matched diode pairs, do you have an idea for making this precision (like .01% control)???

During slew rate limiting, only 2 diodes are on. For example, if the input is rising faster than the capacitor can charge, only D2 and D3 are conducting. Mismatches do not cause slew rate imbalances. Matching is only an issue when slew rate limiting is not occurring. If diodes are mismatched, you then have a DC offset between input and output. Is this what you are referring to? If so, the only suggestion I have is matched diodes, such as CA3039. Current source mismatch will cause slew rate imbalance between positive and negative slopes, and will also cause an offset error.

 

[xmat]

Thanx

Thanx for the tips..

I am also concerned with the amount of control..I 'll have to check in action if it suits my needs well...Anyway thanx a lot.
Really appreciated.

xmat.

 

[Roff]

Re: Thanx

Thanx for the tips..

I am also concerned with the amount of control..I 'll have to check in action if it suits my needs well...Anyway thanx a lot.
Really appreciated.

xmat.

You can control the current sources with a potentiometer or a voltage. Ask and ye shall receive.

 

[Optikon]

Here is a way that does no lowpass filtering on the input. It is a little more complex.

I've attempted to do precision slew control with a similar circuit in the past. It used the same current sources through diodes to a capacitor idea. I couldnt reconcile the fact that the diodes would not equally share the current (and hence error in slew control) this certainly fell apart over temperature. Balancing the voltage on both sides didnt entirely eliminate the issue either. Other than matched diode pairs, do you have an idea for making this precision (like .01% control)???

During slew rate limiting, only 2 diodes are on. For example, if the input is rising faster than the capacitor can charge, only D2 and D3 are conducting. Mismatches do not cause slew rate imbalances. Matching is only an issue when slew rate limiting is not occurring. If diodes are mismatched, you then have a DC offset between input and output. Is this what you are referring to? If so, the only suggestion I have is matched diodes, such as CA3039. Current source mismatch will cause slew rate imbalance between positive and negative slopes, and will also cause an offset error.

I forgot to mention my problem was speed. I had to keep all four diodes biased (and never shut off) to get 100+ Mhz performance out of them with precision. This is where I ran into difficulties. Imagine the operation in the sense of equal bias sharing on all four and then the input signal would unbalance this state of affairs causing a difference current to charge the slew rate cap. Works just as well on discharging. I think I had attempted to even keep the voltages balanced on both sides. But at < .01% accuracy, it quickly pointed to matched diodes for temperature change reasons. I hated the notion of needing matched diodes so I think I scrapped the whole thing for a different method.

But this strays from the original post where this works just fine.. I thought I'd poke your brain about it since I had much frustration trying to get something like this to work at break neck speeds.

 

[xmat]

Re: Thanx

Thanx for the tips..

I am also concerned with the amount of control..I 'll have to check in action if it suits my needs well...Anyway thanx a lot.
Really appreciated.

xmat.

You can control the current sources with a potentiometer or a voltage. Ask and ye shall receive.

Any ideas are always welcome!!

Thanx,

xmat.

 

[Roff]

Re: Thanx

Thanx for the tips..

I am also concerned with the amount of control..I 'll have to check in action if it suits my needs well...Anyway thanx a lot.
Really appreciated.

xmat.

You can control the current sources with a potentiometer or a voltage. Ask and ye shall receive.

Any ideas are always welcome!!

Thanx,

xmat.

What is the frequency and amplitude range of your input? What is the required range of slew rate limiting? What supply voltages do you want to use? Do you need to have a control voltage input, or can you use a potentiometer (i.e., do you want electronic or manual control)?

 

[xmat]

Re: Thanx

Thanx for the tips..

I am also concerned with the amount of control..I 'll have to check in action if it suits my needs well...Anyway thanx a lot.
Really appreciated.

xmat.

You can control the current sources with a potentiometer or a voltage. Ask and ye shall receive.

Any ideas are always welcome!!

Thanx,

xmat.

What is the frequency and amplitude range of your input? What is the required range of slew rate limiting? What supply voltages do you want to use? Do you need to have a control voltage input, or can you use a potentiometer (i.e., do you want electronic or manual control)?

Frequency ranges from 50Hz to no more than 10kHz..
Amplitude range is 100mV to 6.5 Volts
Supply voltage: +-5 to +-9 (dual voltage)
Required range of slew rate limiting: have to check, but I am assuming I have to "filter" signal variations that have 50us to 50 ms occurence time..
Via voltage control input, of course...

Hope I have replied satisfactorily.

Regards xmat.

 

[Roff]

Here's a couple of slew rate limiter implementations. The bipolar version might be easier to get parts for. You could hand-match the trannies, or try to find matched pairs (Q2 and Q4, Q3 and Q5). 1% resistors would also help.
The MOSFET version should somewhat have better offset and +/- slew rate matching, limited only by op amp input offset voltages and resistor matching.
The diode matching only affects offset voltage. You can hand-match the diodes with a multimeter, or you could use a CA3039, if you need low offset voltage. No transistor matching is required.
The op amp was chosen for rail-to-rail input voltage range. You could probably find a substitute.
I haven't built these. They simulate well, and the bandwidth is well beyond 10kHz, even with the slew rate control at 0.1V. Note that for small signals (no slew rate limiting), the diode bridge looks like about 26 ohms divided by the control voltage (Vctrl). At Vctrl=1.0v, it will look like 26 ohms. At Vctrl=0.1V, it will look like 260 ohms. The significance is that the small-signal bandwidth will change as a function of the max slew rate, but it remains well above audio range. You will need to drive the bridge with a source that can handle the load (Rbridge in series with 1nF). Any DC source resistance will cause an offset voltage if the bridge source and sink currents aren't matched.

If you are doing audio glitch reduction, you probably don't need good matching or offset voltage. I'm still not clear on what you are doing. You may have to alter the currents and/or the capacitor to get the results you want. Or maybe slew rate limiting won't really solve your problem. It definitely isn't what you were originally looking into (I can't remember what it was called), which I'm pretty sure requires digital processing (mostly because it needs delay).

 

[xmat]

Great... Thanx.

I'll check and report back if the schematics will be usefull.

xmat.