bizarre behavior, VR refuses to start from int batt, PS or ext batt starts fine

[WastelandTek]

Hoping for some help here guys, this one has me stumped.

My next door neighbor asked me to try and fix this thing.  It's a drag racing tool, basically a few met sensors (temp, bar, hum) and a little micro that calculates your optimum tire pressure based on those readings and a few other input parameters, pretty simple.  He said it wouldn't power up after being dropped.

I put fresh batteries in it, it runs off 4@AA in series and sure enough, nothing, so I did the basic things; cleaning up and reseating the ribbon cable to the keypad and the pin header for the display...still nothing.

OK, time to get serious.  To get it out of the case to start tracing things I had to desolder the battery tabs.  They poke through the back of the case, hefty tabs that solder into big slots on the board, so I got it out and tacked a couple scrap component leads on to those power pads to hook some jumpers to a bench PSU and it boots right up.  Mechanical right? Cold joint or something?  No.

I have it back in the lower case with the battery tabs soldered back on and it will boot off a power supply clipped to those tabs, or off a separate battery 4 pack clipped to those tabs, but absolutely refuses to boot with its own battery case populated.  I can boot it off the external pack and then add the internal batteries, then disconnect the external pack and it will stay on and operate normally, but turn it off and it will not turn back on.

I have traced the problem to a TPS7250Q 5V regulator (datasheet here).  This VR will not start with the batteries in the integrated holder.  It starts every time with batteries clipped on.

I have tried varying the PSU voltage to see if it is over senstitve to that, it is not.  I have tried fresh batteries and somewhat used batteries, both work in the external pack but not in the onboard pack.  I have pushed and tugged on the + and - tabs while it is on trying to find a mechanical issue and there is none.  I have verified voltage on the input pins of the VR when it will not start.

I am left with the puzzling possibility that something about having the battery connected by 20cm of wire vs direct to the board somehow makes a difference.

Please help, I am starting to entertain supernatural explanations here...tearing my hair out.

 

[floobydust]

This thing looks easy to drop, I would check the input and output capacitors to see if they are cracked, as this Vreg can go unstable and oscillate.
Also it doesn't look like the IC can't handle reverse-polarity if the batteries ever went in backwards.

The battery holder might have a bad connection not on the ends (+,- out) but with the middle cells.

[WastelandTek]

This thing looks easy to drop, I would check the input and output capacitors to see if they are cracked, as this Vreg can go unstable and oscillate.
Also it doesn't look like the IC can't handle reverse-polarity if the batteries ever went in backwards.

The battery holder might have a bad connection not on the ends (+,- out) but with the middle cells.

Agree that the design leaves something to be desired, and no, there is no polarity protection I can see, nor is there a real on/off switch, the 3v3 and 5v regs and the micro are always lit...the "off" button puts the micro in a sleep state 

I have ruled out a discontinuity in the battery connections, I can see the battery voltage on the input pins of the VR.

I have looked at the caps, but I will look closer, thank you for the reply.

[ArthurDent]

There is a possibility that when you try to turn it on there is inrush current that causes voltage from the internal battery pack to drop for a few microseconds and the unit senses that and instantly turns off (or not start). If there is any added resistance in the internal battery circuit, that could cause the problem you're seeing, as could an electrolytic capacitor that has opened, as others have said. If there is 5V used internally and the battery supply is only 6V then it wouldn't take much to cause a problem.

Try connecting a capacitor (say in the 100-500 Mf range) across the tabs you have previously connected the PS to and see if that allows the unit to start on the internal batteries. If it does inspect any caps near the battery holder/regulator and all connections. Maybe triggering an A.C. coupled scope set to single sweep on a negatively going pulse connected across the tabs using the various supplies will show the problem.   

[Samogon]

Usually there is battery sense resistor about .1-.2 Ohms, 20sm wire can have such resistance.

[Armadillo]

I would lightly sand the battery contacts off invisible dirt/rust.
Also the battery may be obstructed and not pushing onto the contacts.
Check physically.
Could those be that simple.?

[Cyberdragon]

I would lightly sand the battery contacts off invisible dirt/rust.
Also the battery may be obstructed and not pushing onto the contacts.
Check physically.
Could those be that simple.?

No, he just said he has voltage.

I agree with Sam that it's probably the battery detection circuit. Either a faulty current sense, voltage sense, or switching circuit.

[Armadillo]

No, he just said he has voltage.

I agree with Sam that it's probably the battery detection circuit. Either a faulty current sense, voltage sense, or switching circuit.

Measure Voltage with a high impedance meter example without current loading meant nothing, it could still be phantom voltage over invisible dirt/rust, unless you load it with a external dc load bank over the contacts drawing enough amps of the battery in case to satisfy that the contacts is OK.

Ghost-busting?   

[Cyberdragon]

No, he just said he has voltage.

I agree with Sam that it's probably the battery detection circuit. Either a faulty current sense, voltage sense, or switching circuit.

Measure Voltage with a high impedance meter example without current loading meant nothing, it could still be phantom voltage over invisible dirt/rust, unless you load it with a external dc load bank over the contacts drawing enough amps of the battery in case to satisfy that the contacts is OK.

Ghost-busting?   

No, if it was still connected to the circuit and had any PHANTOM VOLTAGE OF BETWEEN 3 AND 5 VOLTS DIRECT CURRENT it would try to start. If there was a resistive connection there would be pulsing from the circuit trying to start and losing voltage, as mentioned by Arthur.

EDIT to satisfy Armadillo

[WastelandTek]

I would lightly sand the battery contacts off invisible dirt/rust.
Also the battery may be obstructed and not pushing onto the contacts.
Check physically.
Could those be that simple.?

No, he just said he has voltage.

I agree with Sam that it's probably the battery detection circuit. Either a faulty current sense, voltage sense, or switching circuit.

Thanks for actually reading.  Yes, I am poking passives around that VR now.  The really odd thing is that it works with a higher impedance to a power source, and not with a lower impedance, which just seems backwards to me.

[Armadillo]

No, if it was still connected to the circuit and had any voltage at all it would try to start. If there was a resistive connection there would be pulsing from the circuit trying to start and losing voltage, as mentioned by Arthur.

Unless you have read anywhere the magnitude of voltage ever measured exactly at the input pins?
Without knowing the value, what is your basis? No value, no basis in troubleshooting.
You are not going to be toyed around the nose by your customer!
Datasheet says minimum 3V~5V depending on types.
So why don't measure all the 7 pins wrt pin 3?

had any voltage at all it would try to start.

   

NO!

[Samogon]

Cleaning contacts would take less effort than this discussion 

[Armadillo]

Furthermore, the LDO output current around 250mA.
This station only consume so low current?
There may be some others we don't know.
Yeah! why no pictures.?

[xavier60]

A momentary negative voltage on the Output pin with respect to Common can cause 3 pin regulators to latch up. The PCB layout would have to be really bad for this to be the reason.

[WastelandTek]

Cleaning contacts would take less effort than this discussion 

You guys really don't think that was the first thing I did?

Look, the circuit will OPERATE from the integrated holder, just not START from it.  I can put 3 AA in the INT holder, clip the EXT pack onto the batt clips, boot it, install the 4th AA, pull the clips and it stays on.

It just won't start with the batteries in the INT holder.

I will work on getting some pics going, PITA to resize them and all.

[Cyberdragon]

@Armadillo

The OP knows what the correct voltage is, are you assuming he doesn't? He did not need to state it when he said he measured it, he knows what it's supposed to be. Also, a logic controlled power circuit does not know the current capacity of a voltage when it tries to power on, which is what I meant when I said "any voltage at all", I meant phantom voltage, so don't take odd exaggerated statements out of context and gloat, that's being rude. The OP said it's not the battery holder, so it's not the battery holder.

[Seekonk]

Have you actually tried adding a capacitor? I didn't see that mentioned again. That would fit in with the dropping.  I would suspect it going into oscillation.  Has the regulated voltage been looked at?  Is it just gone?  That seems basic.

Just the other day my camera didn't work, screen said recharge batteries. These were just alkaline AA. I was short of batteries an figured I could just replace one or two of the weakest ones to get that one picture I needed. I have been able to do in the past. Measured each one as being over 1.506V.  Put some well used carbon ones in that were only 1/4V and got that picture.  I've never seen camera or batteries act that way before.  Some internal weird bad weld? You can't take anything for granted.

[Armadillo]

@Armadillo

The OP knows what the correct voltage is, are you assuming he doesn't? He did not need to state it when he said he measured it, he knows what it's supposed to be. Also, a logic controlled power circuit does not know the current capacity of a voltage when it tries to power on, which is what I meant when I said "any voltage at all", I meant phantom voltage, so don't take odd exaggerated statements out of context and gloat, that's being rude. The OP said it's not the battery holder, so it's not the battery holder.

I am not assuming. You are.
If you reflect on your own form of response, I would say you deserve what you get, period!.
I merely return back your "NO" "NO". 

[WastelandTek]

a pic, please disregard R18, I had it off to measure, it's good, I'll straighten it up later

[WastelandTek]

here it is operating off the internal holder, having been started with 1 AA out and the ext pack clipped on

[Armadillo]

Have you checked the 2 caps?
ESR and capacitance?
Seekonk suggested adding caps, sound logical.

Please make contact resistance and voltage drop measurement.
Measure from battery point of contact to the VCC IC point.

Also old battery developed higher internal resistance, so in your troubleshooting, you should use all new fresh alkaline batteries.

[WastelandTek]

Have you actually tried adding a capacitor? I didn't see that mentioned again. That would fit in with the dropping.  I would suspect it going into oscillation.  Has the regulated voltage been looked at?  Is it just gone?  That seems basic.

I tried 200uF across the terminals, no go.

I substituted another part for C14, no go.

Do you mean scoping the 5V while it is on?  I detect no oscillation on the output when it fails to start.

[xavier60]

If the regulator is being held off by its Enable pin, maybe this is a normal state and the unit is waiting for a button press to wake it up.
There might be a power button that is supposed to briefly apply voltage to the micro-controller which then enables the regulator.
I realize that this does not directly explain the odd behavior.

[WastelandTek]

If the regulator is being held off by its Enable pin, maybe this is a normal state and the unit is waiting for a button press to wake it up.
There might be a power button that is supposed to briefly apply voltage to the micro-controller which then enables the regulator.
I realize that this does not directly explain the odd behavior.

The "on" button goes directly to the microcontrollers sleep state pin, the VRs are always on if batteries are in...provided the 5V one starts.