LM317 current injector/boosting circuit

Thread Starter

Jake-S

Joined Dec 2, 2020
8
Hello,

I have an application whereby we are sending a 4-20mA signal to an engine controller (engine speed proportional to mA). The test rig has a requirement for throttle blips, which the legacy controller does via a 24V switched output. The new engine controller cannot accept a digital input to allow throttle blips (new engine speed control is via mA signal).

Subsequently I would like to boost the controlling current (which could be between 4-20mA at any given time) to 20mA when the legacy controllers 24V digital output goes high.

I am envisaging using an LM317 but instead of conventionally using a current shunt across the Vo and Vadj pin, instead placing the shunt in the line that will see the 'sum' current.
1677690767966.png

As before, the difference here is that the shunt resistor will feel the total current, and thus the current seen at the engine controller AI would hopefully sit at 20mA when the blip is active.

I was going to size Rshunt as per:
Io = (1.25-1)/R1
With the -1 component due to the forwards voltage of the diode.

The question is.... is use of an LM317 in this configuration suitable? Any comments/ideas also appreciated.

Thanks, Jake
 

ericgibbs

Joined Jan 29, 2010
19,091
hi Jake

Is this what you had in mind, the resistor values need work, also the 24V needs reducing.
E
EG57_ 644.png
 
Last edited:

crutschow

Joined Mar 14, 2008
34,836
Is this what you had in mind, the resistor values need work, also the 24V needs reducing.
How did you get 20mA with only a 0.1Ω shunt resistor?
My simulation shows you need around 25Ω.

Why does the 24V need to be reduced since the LM317 is rated for 40V?

To prevent any possibility of reverse voltage damage, I would add a diode in series with the ADJ connection also.
That will raise the needed shunt resistor to about 50Ω

The pinout numbers in post #1 seem incorrect (below):
1677691908311.png
 

ericgibbs

Joined Jan 29, 2010
19,091
Post your simulation and asc file, we can then compare results.

As always, I try the TS's method he has posted, personally I would not solve the problem using that method.
E
 

ericgibbs

Joined Jan 29, 2010
19,091
Hi,
Your circuit works OK with the added diode, but without the diode, the circuit needs to be fudged.
I did tell the TS his values needed work.
Personally, I think it is preferable to show the TS that his intended circuit will not work as he expects, then make suggestions and changes.

E
Added sim without diode
EG57_ 646.png

EG57_ 647.png
 

crutschow

Joined Mar 14, 2008
34,836
Your circuit works OK with the added diode, but without the diode, the circuit needs to be fudged.
If by "fudged" you mean changing the value of the shunt resistor, of course.

So how did you get 20mA with a 0.1 ohm shunt?

You seem to be following the annoying habit by many posters of ignoring questions asked.
 
Last edited:

Thread Starter

Jake-S

Joined Dec 2, 2020
8
Hi guys,

Thanks for the comments. I will download LTspice and have a play.

I also thought that the 24V trigger feed would not need reducing, as per the LM317 rating.

I assume the reason for the clipping current slightly increasing as I1 increases is because the diode forward voltage increases with current? This gets me questioning whether the diodes are even really required?
Thanks
Jake
 

ericgibbs

Joined Jan 29, 2010
19,091
You seem to be following the annoying habit by many posters of ignoring questions asked.
Hi Carl & Guys,
You must remember the UK is about 5 hours ahead of the USA times.
It was about 7.30pm UK time when I got your question.
As I had been on the site moderating and answering queries since 8am, I was tired, for an 90-year-old guy that's not a bad daily regular effort.

The answer to the question is below, also note this comment I made to the TS.
hi Jake
Is this what you had in mind, the resistor values need work, also the 24V needs reducing.


Again to remind you, I will continue to initially work on the TS's posted circuits and questions, then progress to making recommendations and modifications to his proposal.

hi papa,
Waiting for an answer — that may never come. Kinda like a widow pacing on the widow's walk
I am sure you have more interesting items to attend, too, other than waiting for an answer to my post.:)

As a more general note, I will answer posts and queries in my own good time and in my own way, as most of you already do, that's the way Forums work.

If anyone has an objection to the way I work, so be it.

I sincerely wish you all well and this post is not made with animosity towards anyone.

E
EG57_ 648.png
 

Attachments

Thread Starter

Jake-S

Joined Dec 2, 2020
8
Depends upon how much voltage the LM317 output sees from the 4-20mA signal when the LM317 is off.
Since the mA output device has a 24V supply I dont think the Voltage could ever exceed this if there was an open circuit condition.

I was thinking I could use an SPDT solid state relay (with the common connected to the engine controller) to switch the control current between:
  1. The variable 4-20mA current source
Or
  1. A LM317 circuit controlling to 20mA

My concern with this however was whether any damage could be caused to the analog output device in the rig controller, since it will periodically go open circuit and thus ramp up to Vcc to attempt to drive its setpoint current. I could work around this with a DPDT relay and switch the 4-20mA line to a load resistor... but I may be overcomplicating this now!
 

ericgibbs

Joined Jan 29, 2010
19,091
Hi @Jake-S
He is concerned about the reverse voltage the LM317 will 'see' when the the 24V input is low.

Ideally with a 250R load on the 4-20mA that voltage will be a maximum of +5V, when the 4-20mA line is at 20mA from the driver.

E

Update: the two existing diodes should block any reverse voltage on the LM317
If considered necessary, a further reverse connected diodes could be added across the LM317, back to the 24V source
For open circuit protection on the 250R lines, you could add a 12V Zener diode across the line,
EG57_ 651.png
 
Last edited:

Thread Starter

Jake-S

Joined Dec 2, 2020
8
Thanks for the feedback Eric.

I like your simplified circuit, however I wanted to make a system that would work regardless of the load resistance (r2), since this can change across devices.

I like the idea of having a zener so that the driver still sources current during open circuit! That being said I think the drivers are fine to be left in the open circuit state, it was more the repeated transients I was concerned about with my comment over the solid state relay idea.
 

ericgibbs

Joined Jan 29, 2010
19,091
Hi Jake,
Could you sketch out your SSR idea and what level of transients you are expecting, we could run some simulations to check their effect.

E
 

ericgibbs

Joined Jan 29, 2010
19,091
Hi Jake,
This sim shows the effect of changing load resistance from 250R to 1000R in 250 steps.
As you can see, other changes to the circuit would be needed to keep the 20mA pulse constant.
E
EG57_ 655.png
 
Top