OK, it's not 10%.
Then why not rectify it with a diode and a capacitor?
(The problem isn't very well defined.)
-- John Larkin Highland Technology, Inc Science teaches us to doubt. Claude Bernard
Yes, that's pretty much an ideal diode and cap: OpAmp(D),RC -> Comp-(10V),Comp+(8V) -> OpenCollectorDriver
That's not very well defined either.
-- John Larkin Highland Technology, Inc Science teaches us to doubt. Claude Bernard
10% duty cycle 1KHz signal, area of interest at 6V and 9V +/- 1V. Other duty cycles and voltages upto 12V may be present and rejected.
Is that clear?
Accuracy?
Duration/decay time constant?
Source impedance?
Measurement device?
Scale of production?
Why not rectify it with a diode and a capacitor?
-- John Larkin Highland Technology, Inc Science teaches us to doubt. Claude Bernard
10%
100ms
1K ohm
Lm358 and Lm393
Low
The Op Amp bring it closer to the source voltage without worrying about the variable diode drop, as well as isolating the circuit from the input.
It's good enough for something called brainstorming. You might try it some time. Lots of people here talk about using it. It's very popular. What do you have to lose? I recommend it.
-- Rick C. -++ Get 1,000 miles of free Supercharging -++ Tesla referral code - https://ts.la/richard11209
om
form.
er/
ak DC voltage of an
do you want to
escriptive.
.t would get confused with 10% 9V and 15% 6v.
Not completely. What does "rejected" mean? You haven't explained exactly how you want the outputs to behave. Do you want two outputs that detect th
voltage is 7.5V neither output is active? If the duty cycle is not 10% sh ould neither output be active?
It's interesting watching the analog approach get more complex as it is ref ined. I'd like to see a BOM.
-- Rick C. +-- Get 1,000 miles of free Supercharging +-- Tesla referral code - https://ts.la/richard11209
e:
m.com
e:
veform.
eter/
peak DC voltage of an
or do you want to
descriptive.
al.
it would get confused with 10% 9V and 15% 6v.
y how you want the outputs to behave. Do you want two outputs that detect
he voltage is 7.5V neither output is active?
Yes, only the prefined voltage ranges are accepted.
The voltage is more important in this case, even if the duty cycle is off. I was just bringing up the point that averaging is not appropriate, but pe ak detection. Anyway, averaging is not easy to do in analog, peak detectio n is simple enough.
efined. I'd like to see a BOM.
Lm358 + 2 x Lm393
ote:
com
il.com
ote:
waveform.
tmeter/
e peak DC voltage of an
, or do you want to
ry descriptive.
gnal.
o, it would get confused with 10% 9V and 15% 6v.
tly how you want the outputs to behave. Do you want two outputs that detec
the voltage is 7.5V neither output is active?
. I was just bringing up the point that averaging is not appropriate, but peak detection. Anyway, averaging is not easy to do in analog, peak detect ion is simple enough.
refined. I'd like to see a BOM.
So the circuit has two outputs from four comparators. Going to use wire OR s to get a high only when within each window?
There will be a number of passives and don't you need a voltage reference? Or will you use the power supply?
Can't compete with the price for sure but for low volumes $0.50 is not much of an issue. I didn't know you could get an op amp/comparator that cheap. But the MCU will be a lot less space on the PCB and have fewer components overall. The programming is just part of the test process when I build bo ards. I usually design a simple test fixture, but that depends on the prod uct.
What else is on this board?
-- Rick C. +-+ Get 1,000 miles of free Supercharging +-+ Tesla referral code - https://ts.la/richard11209
ote:
com
il.com
ote:
waveform.
tmeter/
e peak DC voltage of an
, or do you want to
ry descriptive.
gnal.
o, it would get confused with 10% 9V and 15% 6v.
tly how you want the outputs to behave. Do you want two outputs that detec
the voltage is 7.5V neither output is active?
. I was just bringing up the point that averaging is not appropriate, but peak detection. Anyway, averaging is not easy to do in analog, peak detect ion is simple enough. Averaging is dead easy in anaolog - it's called "low pass filtering". It ca n be a bit slow.
refined. I'd like to see a BOM.
Plus power supplies, lots of resistors, the occasional capacitor ...
Waht does seem to be missing is an an explicit and more or less complete sp ecification of what the the circuit does. Why is does it is frequently usef ul too.
-- Bill Sloman, Sydney
More refinement should simplify a design. The popular concept of "over-engineered" is backwards.
This might work as your peak detector. It might be good enough without D1+R1, especially if you account for the drop of D2 downstream.
Incidentally, averaging is really easy: one RC.
-- John Larkin Highland Technology, Inc Science teaches us to doubt. Claude Bernard
:
om
l.com
mail.com
e waveform.
V:
oltmeter/
?the peak DC voltage of an
et, or do you want to
very descriptive.
signal.
So, it would get confused with 10% 9V and 15% 6v.
.actly how you want the outputs to behave. Do you want two outputs that det
if the voltage is 7.5V neither output is active?
ff. I was just bringing up the point that averaging is not appropriate, bu t peak detection. Anyway, averaging is not easy to do in analog, peak dete ction is simple enough.
is refined. I'd like to see a BOM.
ORs to get a high only when within each window?
Yes, actually the ltc2304 can do it in single chip, but expensive.
? Or will you use the power supply?
Just resistor divider voltage reference: 2K,1K,1K,1K,1K,6K.
ch of an issue. I didn't know you could get an op amp/comparator that chea p.
Brought 50 lm358 and 100 lm393 for less than $10, including shipping from C hina. Hopefully, they are open now. I can burn out lots of them doing pro totype.
s overall. The programming is just part of the test process when I build b oards. I usually design a simple test fixture, but that depends on the pro duct.
Space is not a problem. If sale volume is high enough, could do micro.
Power supply and relays.
D1+R1, especially if you account for the drop of D2 downstream.
This is pretty much same as our ideal diode circuit, but just one diode. Don't understand why you want D2 and bring back the voltage drop problem. Op amp eliminates the voltage drop error across D1.
te:
.com
ed.
exactly how you want the outputs to behave. Do you want two outputs that d
o if the voltage is 7.5V neither output is active?
off. I was just bringing up the point that averaging is not appropriate, but peak detection. Anyway, averaging is not easy to do in analog, peak de tection is simple enough.
t is refined. I'd like to see a BOM.
e ORs to get a high only when within each window?
I can't find that device on ADI's web site... it is mentioned in some Russi an viewgraphs.
ce? Or will you use the power supply?
much of an issue. I didn't know you could get an op amp/comparator that ch eap.
China. Hopefully, they are open now. I can burn out lots of them doing p rototype.
$25 would buy plenty of MCUs and they'd be reusable, not likely to get burn t out. China is already on Monday, so yeah, they're open, well, shortly.
nts overall. The programming is just part of the test process when I build boards. I usually design a simple test fixture, but that depends on the p roduct.
The two diodes connected that way eliminate the saturation of the op amp ou tput which allows a response to shorter pulses which is not needed at all i n this design. It also mucks up the diode voltage drop compensation since the two diodes will have different currents when charging up the cap. That was the main variation in the many peak detector circuits I saw, compensat ing for the diode voltage vs. keeping the op amp out of saturation... to so me degree. None of them manage it perfectly. This is the first one I've s een that slacked off on the diode drop compensation.
Analog is messy and requires a lot of iteration to get right. It often als o needs various... circuitry to deal with detail issues. The op amp is the first order of this. The diode has a voltage drop, so add the op amp to f ix that. The op amp saturates and is slow to respond to the leading edge o f the pulse, so add more circuitry to fix that, etc.
You probably can do without the op amp actually. The diode drop can be red uced by using a Schottky diode and even then if the end current is not larg e it will be near the low end of it's range. With your voltages being larg e compared to the diode drop it should be no problem to compensate for it i n the resistor dividers. You've got a 10% window and 1% resistors. Accoun ting for a quarter volt diode drop should be pretty trivial.
BAT54J VF forward voltage[1] IF= 0.1 mA--240mV IF= 1 mA--320mV IF= 10 mA--400mV IF= 30 mA--500mV IF= 100 mA--800mV
-- Rick C. ++- Get 1,000 miles of free Supercharging ++- Tesla referral code - https://ts.la/richard11209
:
rote:
il.com
V.
cted.
d exactly how you want the outputs to behave. Do you want two outputs that
So if the voltage is 7.5V neither output is active?
is off. I was just bringing up the point that averaging is not appropriate , but peak detection. Anyway, averaging is not easy to do in analog, peak detection is simple enough.
it is refined. I'd like to see a BOM.
ire ORs to get a high only when within each window?
sian viewgraphs.
Sorry, tlv2304. Op amp + comparator in single chip.
te:
mail.com
1V.jected.
ned exactly how you want the outputs to behave. Do you want two outputs th
? So if the voltage is 7.5V neither output is active?
e is off. I was just bringing up the point that averaging is not appropria te, but peak detection. Anyway, averaging is not easy to do in analog, pea k detection is simple enough.
as it is refined. I'd like to see a BOM.
wire ORs to get a high only when within each window?
ussian viewgraphs.
That would not accomplish much considering you need four comparators.
If you are running the relays off the same power as the rest of the circuit you might need to filter the reference resistors pretty well to avoid glit ches and/or oscillations. Circuit activates causing relay to trip which su rges the power supply drooping reference voltage causing comparator to drop out cycling around a few times.
So add some caps to your BOM.
I just would never try to do this with analog devices this way when it is s uch a simple design for an LDO, an MCU and two transistors - $1 and you are done. I expect for not much more you can buy MCU modules that solder to y our board. In fact, I bet you can find relay boards with an MCU ready for you to add your programming.
Why roll your own board when you can buy hardware that will do this?
-- Rick C. +++ Get 1,000 miles of free Supercharging +++ Tesla referral code - https://ts.la/richard11209
rote:
@gmail.com
er
/- 1V.
rejected.
ained exactly how you want the outputs to behave. Do you want two outputs
se? So if the voltage is 7.5V neither output is active?
cle is off. I was just bringing up the point that averaging is not appropr iate, but peak detection. Anyway, averaging is not easy to do in analog, p eak detection is simple enough.
x as it is refined. I'd like to see a BOM.
se wire ORs to get a high only when within each window?
Russian viewgraphs.
it you might need to filter the reference resistors pretty well to avoid gl itches and/or oscillations. Circuit activates causing relay to trip which surges the power supply drooping reference voltage causing comparator to dr op out cycling around a few times.
Solid state relay takes around 25mA to activate, should not cause to much r ipple.
cuit you might need to filter the reference resistors pretty well to avoid glitches and/or oscillations. Circuit activates causing relay to trip whic h surges the power supply drooping reference voltage causing comparator to drop out cycling around a few times.
ripple.
Ah, the big buck relays. Low voltage/low current switched signal? I know for power line apps magnetic relays are much more cost effective. Obviously the load is not on the same supply or the relay would not be needed.
I want to make a control for my 240 volt, 30-50 amp household devices. No way I'm using a solid state relay for that. Even a magnetic relay goes up in price significantly from a 30 amp unit.
-- Rick C. ---- Get 1,000 miles of free Supercharging ---- Tesla referral code - https://ts.la/richard11209
ircuit you might need to filter the reference resistors pretty well to avoi d glitches and/or oscillations. Circuit activates causing relay to trip wh ich surges the power supply drooping reference voltage causing comparator t o drop out cycling around a few times.
ch ripple.
w for power line apps magnetic relays are much more cost effective. Obvious ly the load is not on the same supply or the relay would not be needed.
o way I'm using a solid state relay for that. Even a magnetic relay goes u p in price significantly from a 30 amp unit.
I only need around 120V 6A. It's less than $5.
You want to detect peaks in the 6V region reliably and in the 9V region too. Power-supply is +12V DC
I still have many questions:
How close to 12V is the supply voltage? how much ripple?
You mentiond 10% PWM may increase to 15%, will it ever go above 90% or below 10%? (and still require correct output) Is there an output signal requirement if the PWM stops? if so What constitutes a stop ? do we need to ignore runt pulses?
Your candidate circuit at the beginning was AC coupled. Is your signal AC coupled or DC coupled? Do you have output signal requirements for out of range signals? I'm assuming 7-8V is treat as either 6V or 9V, but what about for below 4V? what about for for above 11V? How fast must the circuit respond to a change in signal amplitude? 100 microseconds? 100 milliseconds? Preferred output is current sinking or sourcing? logic level? (which)
Can I draw 20mA from the signal line or it it high impedance?
-- Jasen.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.