TVR on a line driver

[Scarr]

Hi All,

I have a line driver EL7232, I want to protect the output A (pin 7) from voltages > 6v to protect the output from incoming high voltages, the EL7232 is supplied from a 5v regulator and I have tried putting a 1V5KE6V8A TVR on output but when I activate the output the MCU resets, probably a really basic error but I cannot see why this would happen.

Thanks

Steve

 

[ronsimpson]

diode in backwards?
Why not a diode from output to +5V?

 

[Scarr]

Hi Ron,

No, diode (TVR) correct way around.

More info - I have programmed the MCU to flash RED on power up and green when output is activated, by "output activated" I mean +3,3v on pin 1 of EL7232 + a low on "A IN" pin 2. it's at this point the MCU resets, If I remove the TVR it pulses fine, time and time again!

Also I have a polyfuse on the output line near the TVR but I cannot see how this would effect things, the reset happens both when nothing is connected to the output or the output wire

Steve

P.S. Not sure what you mean by "Why not a diode from output to +5V?"

 

[Tony Stewart]

INcoming high voltages is a concern and sounds like a very inductive load such as a motor or perhaps a large capacitive load is causing a Voltage dip and triggers the POR (reset)

Your driver is rated as follows;

• 2A Peak drive
• Low, matched output impedance 5Ω (+/-50% est.)

For practical design, Supply ESR = 1% load ESR for 1% voltage regulation. (1) Never forget this. This includes all caps, and inductance, drivers and regulators

​

ESR is the effective series resistance of all components where ESL is the inductance (1nh/mm).

Let's assume your load is highly reactive with low ESR such as a solenoid or motor or a capacitor bank with low ESR.

Likely your problem is not the TVS diode you added, but rather your load current thru the driver causes a voltage drop to trigger the power-on-reset cct. (POR)
My guess from your lack of info is this is your problem.


In fact it is useless for clamping low voltage reactive loads. It is intended for ESD and lightning transients.

From the datasheet, 1V5KE6V8CA Vth= 6.45 to 7.14 @10mA , which means at 7.14V or 2.14 above 5Vrail it only absorbs 10mA or effectively 2.14/10mA or 214 Ohms load.
Whereas 1A, 1W diodes from output to both supply rails ( reverse biased) will provide an ESR ~ 1Ω for 1W rating at saturation of <1V drop.

Solutions:

Analyze your load current and supply dip from impedance ratio of ESR or Rs, series resistance of the reactive load.
Add 1% of this load in low ESR cap across the driver supply with a RC time constant at least 5x the duration of peak current.
If reactive load, p ( e.g. motor) put protection diode across inductive load not the driver to isolate clamp current to just the loop around the load so as not to spike ground voltage from V=Ldi/dt from wiring or similarly spike Vcc as a high side switch to 5V or driving a low ESR capacitance Ic=C dv/dt .

***edit**
- having just read your reply on LED or no load causing the problem with TVS across output, keep in mind the TVS has a capacitive load at least 10nF with 5V bias and that is very low ESR since it has a power handling capacity of 1500W so ESR will < 1mΩ when saturated . when not conducting, it is just a low ESR cap of 1mΩ .
My Rule of thumb for semiconductors is ESR is ~ 1/Watt rating.

Solution

1. use a very low ESR electrolytic Cap sufficiently large to drive load or 4.7uF min. across driver Vcc to gnd
2. remove TVS and use a low capacitance high speed diode ( rated for load current) from output to supply rail across inductive load, not the driver in reverse bias. with normal 0~5V range. With switching to Gnd causes spike current add another reverse biased high speed diode from output to Gnd rail.

Summary: from data you have provided it is simply lack of decoupling or ground plane or between supply and driver IC. Also consider low ESL ground track that does coupling to nearby POR signal.

What is your intended load anyways?

 

[Tony Stewart]

from reading your other questions, it appears you want to drive a high power White LED from 5V with a 3~5 Ohm driver so you get 1~1.5Amp and possibly help your wife.
Depending on pulse duration, the regulator and ESR and layout ESL MUST be much lower than ESR of LED which is around 1/3 Ω @ at 3W.

Using my Rule of Thumb, ESR source = 1% of load or 3 mΩ may seem like overkill, but then considering the EL7232 has a switching time of 20ns with a 10,000 pF TVS diode load with near zero ESR... Ic=Cdv/dt = 1e-8F*5V/2e-8s = 2.5A but since driver ESR is 3~5Ω, the current transient on the TVS will be limit to 1.5A which is effectively "NOLOAD" in your assumption, but in fact is the same as driving a short circuit for 20ns or driving a 3W LED.

Thus a better understanding of ESR and load ratios of source to load will help you understand why you need to decouple better with low ESR cap at source and low stray inductance of wiring that can couple into signals on MCU. ( suggest you use shielded or twisted pair and good pulse load current loop survey of driver current on layout and consider ways to reduce ESR,ESL using flat strips, or braided wire. with supply twisted pair as well supported by low ESR cap maybe <<50 mΩ

 

[Tony Stewart]

if you rely on driver to provide voltage drop to LED using Vcc to anode, consider the VI * duty cycle power dissipation on the driver when you drive the LED.

IMHO, it is better to use White LEDs at rate current = 2.9V using a 3V Lithium battery as large capacitor for decoupling low cost with high current spikes with very low ESR then add a 0.1uF low ESR ceramic and 50 mΩ Mosfet to drive the LED with low voltage drops across all devices for PWM. If running continuously and hot, then a SMPS CC LED driver is best to prevent thermal runaway.

Alternatively better if you want more LED's , put in series just below decent power supply ( e.g. x4 12V on PSU and regulate pulse duty cycle to maintain optimal average current which may be between 90 and 100% depending on heatsink Tj rise and thus very efficient.

In future please give photos of layout, and more info on initial question.

 

[Scarr]

Ignore the post below, It's fixed, not sure why, can only think it was a bad connection somewhere, all I did was go off and have a cold drink, come back and re-soldered a few connections & tested it, and it works!

Hi all,

Well it was the lack of caps causing the MCU reset and I have taken the advice of NOT using a TVR but instead I chose a crowbar circuit. This time I have provided everything, even a picture of the circuit (all be it a few component thrown together very quickly to test it). this circuit does not behave as I think it should and that is, using a bench PSU at 2v the polyfuse is activating as if there is a dead short and I cannot understand why, from what I understand this should be OK upto 6.8v approx then the SCR should conduct as the zener breaks down and activate the polyfuse.



P.S. please forgive the nasty solder braid as a earth wire but it was to hand and I hope this time I have given enough information. also I will be putting a cap in there but wanted to get it working first.

Steve

 

[Scarr]

Hi all,

I am beginning to despair with electronics as I hate keep asking what I feel are stupid questions! but here I go again.

The EL7232 is a push pull line driver capable of sourcing / sinking 2A between 4.5v and 16v but the inputs can be operated from 3.3v

Here is what I do:

1A 5v regulator MC7805CDTRKG
EL7232
Atxmega running at 3.3v
The 5v regulator supplies only the EL7232, I have a separate regulator for the MCU



Set pin 1 (3-STATE for line A) high at (3.3v)
Cycle PIN 2 high and low at 3 second intervals, this should drive the output, pin 7, high / low (inverted)

To test the "PULL UP" I connect the output it to ground via 6ohm resistor, I see the output drop from 5v to 1.73v and it see it drawing 290ma approx
To test the "PULL DOWN" I connect the same resistor to the output of the 5v regulator and I see 3.54v and drawing again 290ma approx.



So here are my questions

• Why does the EL7232 not pull harder and keep the line high / low as it cycles
• Why do I only see 290ma through the resistor when it should be 833ma approx (5v/6ohm = 0.8333)

Is this something to to with VCC to the EL being 5v and the pins being driven at 3.3v?

I hope someone can help and please please please do not get pissed off with me, I do spend a lot of time trying to figure it out before thinking of posting, maybe too long!

Steve

 

[Tony Stewart]

Hi all,

I am beginning to despair with electronics as I hate keep asking what I feel are stupid questions! but here I go again.

The EL7232 is a push pull line driver capable of sourcing / sinking 2A between 4.5v and 16v but the inputs can be operated from 3.3v

Here is what I do:

1A 5v regulator MC7805CDTRKG
EL7232
Atxmega running at 3.3v
The 5v regulator supplies only the EL7232, I have a separate regulator for the MCU



Set pin 1 (3-STATE for line A) high at (3.3v)
Cycle PIN 2 high and low at 3 second intervals, this should drive the output, pin 7, high / low (inverted)

To test the "PULL UP" I connect the output it to ground via 6ohm resistor, I see the output drop from 5v to 1.73v and it see it drawing 290ma approx
To test the "PULL DOWN" I connect the same resistor to the output of the 5v regulator and I see 3.54v and drawing again 290ma approx.

View attachment 93011

So here are my questions

• Why does the EL7232 not pull harder and keep the line high / low as it cycles
• Why do I only see 290ma through the resistor when it should be 833ma approx (5v/6ohm = 0.8333)

Steve

All you need to apply is Ohm's Law with an understanding of each driver output impedance and the power dissipation temperature rise.
However, you do need to provide a full schematic with photo of layout for details.

The 3 terminal regulator is a regulated Emitter Follower with current sense with OCP and OTP internal.
The EL7232 is a logic level MOSFET half-bridge switch with 3~5 Ohms RdsOn which will not draw 2A under the circumstances. Use a better Half-Bridge design to exceed the load and transient load and supply currents with minimal drop ( ie <2% ), keep in mind the ESR of load caps and Diodes.(LEDs) can be computed or measured.

I suspect your load ESR is very low with ? uF of capacitance. therefore reduce this and put before switch only using small caps on MCU. ( <=0.01)
LOad surge on pulsed voltage can be analyzed once entire design is known.

Use Ohm's Law for voltage division using ESR's or supply more info and ask specifics.