Roomba 500 R52

The very latest news and updates for the iRobot Roomba robotic vacuum, the Scooba Robotic Washer and the Dirt Dog workshop sweeper. All discussion and troubleshooting questions go here.

Roomba 500 R52

Postby blevine » September 12th, 2012, 1:52 pm

I'm having problems charging my Roomba 530.
I'm thinking it might have something to do with detection of the thermistor in the battery.
Unconnected this has a value around 10k which seems fine so I'm thinking it has something to do with the detection circuit.

If I apply external power to the Roomba then it always enters charging mode, regardless of whether a battery is connected or not. However, the charging fets do not get activated and the battery does not charge.

This seems to be similar to viewtopic.php?f=1&t=15704&p=103988

In either case, with a battery connected or not, according to the SCI output the tenths-deg-C value is 797 which looks pretty high.

Looking at the board I notice my R52 resistor looks singed i.e. slightly brown underneath. This resistor in right in front of the thermistor connection nearest VBAT.

Can anyone give me a resistor value for R52 ? I can't make out any resistor marking. I'm getting a value of 10k from this resistor.

Thanks
blevine
 
Posts: 12
Joined: September 12th, 2012, 4:57 am

Re: Roomba 500 R52

Postby TechGuy » September 12th, 2012, 4:20 pm

R52 on my 560 has 470 marking on it. It is a 47 ohm resistor.

Off-Topic
How to read SMD resistors
First 2 digits are values
The 3rd and the 4th digits are number of zero appended to the end of the first 2 digits
470 = 47 ohms
471 = 470 ohms


If you have an Android phone, you can download a free app from Google Play called ElectroDroid. This is a handy tools for quick reference.
Charging battery directly: 400 Series, 500 Series,.H-Bridge Repair How to Desolder
TechGuy
Robot Addict
 
Posts: 1585
Joined: February 17th, 2009, 11:52 pm

Re: Roomba 500 R52

Postby blevine » September 12th, 2012, 6:01 pm

Thanks for the info.
My value for R52 is actually 1.6k, still very far from your 47R.
I'm also getting 3 ohm when measuring the channel of D19 (which has markings U8 A05 which I think is a BSR14 NPN transistor). This is with the battery pulled and no external power.

I'm trying to trace out the path for the thermistor detection circuit.
blevine
 
Posts: 12
Joined: September 12th, 2012, 4:57 am

Re: Roomba 500 R52

Postby Gordon » September 13th, 2012, 2:42 am

blevine wrote:...I'm also getting 3 ohm when measuring the channel of D19 (which has markings U8 A05 which I think is a BSR14 NPN transistor). ...
I think that A05 SOT-SMD will turn out to be a dual diode. I can't recall if A05 is cathode to cathode or the diodes are in anode to cathode series. I can look it up tomorrow PM.
Gordon
Robot Master
 
Posts: 4304
Joined: April 6th, 2005, 2:02 am
Location: Santa Ynez, CA USA

Re: Roomba 500 R52

Postby blevine » September 13th, 2012, 5:04 am

Based on your post here:
viewtopic.php?f=1&t=12229&start=20
that makes sense.

Image

I'm pretty sure pin 2 isn't connected to anything on the main board.

Based on my measurements the whole package is shot. I'm measuring 3 ohm between pins 2 and 3. With 10k ohm between pins 1 and 3 and between pins 1 and 2. I can't see any diode-like behaviour on the pins.
blevine
 
Posts: 12
Joined: September 12th, 2012, 4:57 am

Re: Roomba 500 R52

Postby Gordon » September 13th, 2012, 10:40 am

blevine wrote:...I'm pretty sure pin 2 isn't connected to anything on the main board. ...
I notice the AO5 being used that way quite often. When the circuit being studied gets sorted out, the function provided is voltage limiting / clipping.

The R52 failure is really odd / rare, since Roomba applied bias current is very little (thinking less than 1 mA) through the thermistor. On that basis I am tending to think this 530 may a previously owned Roomba whose mobo was abused by another person.
Gordon
Robot Master
 
Posts: 4304
Joined: April 6th, 2005, 2:02 am
Location: Santa Ynez, CA USA

Re: Roomba 500 R52

Postby blevine » September 13th, 2012, 10:52 am

I bought it off ebay last week as "for parts/not working". I asked the seller about the problem and he said:
"All I can tell you is that I put it on charge one night and in the morning it would not work."

Seems suspicious...
blevine
 
Posts: 12
Joined: September 12th, 2012, 4:57 am

Re: Roomba 500 R52

Postby TechGuy » September 13th, 2012, 12:26 pm

blevine wrote:...
"All I can tell you is that I put it on charge one night and in the morning it would not work."

Seems suspicious...

If water gets in from the top (Pets use it as toilet), it will kill it...
Charging battery directly: 400 Series, 500 Series,.H-Bridge Repair How to Desolder
TechGuy
Robot Addict
 
Posts: 1585
Joined: February 17th, 2009, 11:52 pm

Re: Roomba 500 R52

Postby blevine » September 19th, 2012, 6:53 pm

Pending further verification, I seem to have fixed the problem.

Initially I only replaced R52 with a replacement 47 ohm resistor. However, this did not solve the problem.
Next, I removed (entirely...) D19.
I now have a charging current of 250mA for a battery already charged to 16V. This looks promising.

However, can someone give me a diode reading for the pins of D19?

Pin 2 of D19 is connected to ground and pin 1 connected to R52. I'm wondering if this is a TVS diode but can find no reference to a package with the marking 'A05 U8'. If it is TVS then I'll need to work out if it's uni or bidirectional.
blevine
 
Posts: 12
Joined: September 12th, 2012, 4:57 am

Re: Roomba 500 R52

Postby TechGuy » September 19th, 2012, 7:26 pm

The charging current should be 1.2A - 1.3A during the initial fast charging mode.
Your Roomba will give you charging error 5 if the charging current is below 400mA for over 30 minutes.
Do you get charging error 5 after charging for 30 minutes?
Charging battery directly: 400 Series, 500 Series,.H-Bridge Repair How to Desolder
TechGuy
Robot Addict
 
Posts: 1585
Joined: February 17th, 2009, 11:52 pm

Re: Roomba 500 R52

Postby Gordon » September 20th, 2012, 3:13 am

blevine wrote:Pending further verification, I seem to have fixed the problem.
I agree with TechGuy: your Roomba is not charging properly.
Initially I only replaced R52 with a replacement 47 ohm resistor. However, this did not solve the problem. ...Next, I removed (entirely...) D19.
Not a good idea to power up Roomba with D19 disconnected. My interpretation of its utility is it prevents the temperature signal voltage heading to the MCU from going negative by more than half a volt. iRobot EEs use those clippers on many lines heading to MCU ports.
...
Pin 2 of D19 is connected to ground and pin 1 connected to R52. I'm wondering if this is a TVS diode but can find no reference to a package with the marking 'A05 U8'. ...
My reference for A05 pinout is not in agreement with the numbers you have stated. You should visit Vishay Semiconductors and search its site for BAV70-V (also search for the A7 dual diode BAV99V, since there are plenty of them planted on R3's main_PCA).

Somewhere along the line I tend to recall that the diodes in these dual-diode packages were claimed to be equivalent to the 1N4148 signal diode

Vishay numbers pins on A05 & A7 the same way a SOT-23 package is pinned if a transistor die is in the package.

Over the last few days I have gotten a pretty good handle on this thermistor biasing circuit, and on the signal path to the MCU. Rather than labor with making a clean schematic diagram I thought it might be quicker to type it an ASCII diagram. I must attach the ASCII portion as an image, however the following details will be needed to flesh out resistance values and other meta data not included in the ASCII form:
====================================== DETAILS ==========================================

COMPONENTS:

C74(-16,57){SMD,ceramicBypass}.

D19(d1)K(-13,57){DualDiode-A05,SOT23}.
D19(d2)(-13,57){DualDiode-A05,SOT23, d2 not used}.

R52(-9,57){SMD,470,47ohmsNOM}.

R316(-12,53){SMD,01C,10kohmsDMMins}.

R317(-10,53){SMD,898,8220ohms}. <corrected 120921, WAS 1.8k ohms>

R319(14,54){SMD,470,46ohmsDMMins}. <added 120920>

R461(49,-3){SMD,102,1kohms NOM}.

U8(1,81)8{Out3,LM324quadOA,14pinSOIC}
U8(1,81)9{IN3(-)LM324quadOA,14pinSOIC}
U8(1,81)10{IN3(+)LM324quadOA,14pinSOIC}


POWER & GROUND:

"Node+5VREG_switched" provides bias current to the thermistor.
*/ Node is 0V in STBY, but +5V in ClnRdy & Charging-Mode states /*
*/ The width of this track is about 0.75 mm and runs to many places on the board. /*

SYS_GND vs. mobo "GND":

Refer to R3's system current shunt resistance composed of SMD power
resistors R235 & R257 wired in parallel. They are located to the right
(Roomba's right) of the battery terminal labelled "GND" = J14.

The left ends of the shunt connect to GND, and it connects to the battery
terminal marked (-).

The right ends of the shunt connect to what I call "SYS_GND", or sometimes
"SYS_RTN" because all of Roomba's subsystems return their individual
currents to the right ends of the shunt. The summation of those currents pass
through (in concept) the shunt on its way back to the battery, and the
total is processed by the MCU & firmware to reveal temporal system current.

SHORTHAND NOTATION: {too late at night to include it -- I'll wait until asked about my notations}
Attachments
in_lieu_of_schematicDWG.jpg
Last edited by Gordon on September 21st, 2012, 6:15 pm, edited 2 times in total.
Gordon
Robot Master
 
Posts: 4304
Joined: April 6th, 2005, 2:02 am
Location: Santa Ynez, CA USA

Re: Roomba 500 R52

Postby blevine » September 20th, 2012, 4:25 pm

TechGuy and Gordon, thank you for your continued support with this problem.

Clearly I jumped the gun last night with my 'it's fixed!' post.

Not a good idea to power up Roomba with D19 disconnected.

I've replaced D19 with an equivalent diode.

My reference for A05 pinout is not in agreement with the numbers you have stated

Sorry, I was referring to an incorrect drawing of the package. Your numbers match up precisely with my main board.

With D19 in place I now have a charging current (as reported over SCI) of 1156mA. Also the temperature measurements seem reasonable, initially reporting as 200 and moving up as charging progresses. Peaking around 320.
The robot also entered trickle mode with a current of 55mA, and the LEDs turn green.

Gordon, to add to your info R319 is marked as '470' which is 47 ohm nominal.

Is there anything further I should be measuring to check the health of the charging circuit?
blevine
 
Posts: 12
Joined: September 12th, 2012, 4:57 am

Re: Roomba 500 R52

Postby TechGuy » September 20th, 2012, 6:57 pm

Congratulation, The fast and trickle charging currents are good.
Charging battery directly: 400 Series, 500 Series,.H-Bridge Repair How to Desolder
TechGuy
Robot Addict
 
Posts: 1585
Joined: February 17th, 2009, 11:52 pm

Re: Roomba 500 R52

Postby Gordon » September 21st, 2012, 1:47 am

blevine wrote:TechGuy and Gordon, thank you for your continued support with this problem.
Welcome!
...Gordon, to add to your info R319 is marked as '470' which is 47 ohm nominal.
Thanks for catching that data list omission. That was due to posting too close to midnight -- at that late hour I tend to not check my work thoroughly!
Is there anything further I should be measuring to check the health of the charging circuit?
Well, TechGuy covered that question, but I am interested in why renewing blown R42 & D19 components overcame this false charging start-up that you were facing in your OP:
blevine wrote:...If I apply external power to the Roomba then it always enters charging mode, regardless of whether a battery is connected or not. ...
Very mysterious!
Gordon
Robot Master
 
Posts: 4304
Joined: April 6th, 2005, 2:02 am
Location: Santa Ynez, CA USA

Re: Roomba 500 R52

Postby blevine » September 21st, 2012, 5:59 am

Just thinking out loud I'd say most of the problem I had was to do with D19.

I had said, with regard to D19, measurements of:
... 10k ohm between pins 1 and 3 and between pins 1 and 2. I can't see any diode-like behaviour on the pins.


I don't think that 10k value of D19 is accurate and I'd guess the diode broke down completely with any significant voltage across it. As a result, the midpoint of R317 and R52 was close to 0V.

With a battery connected, the output of the junction of R317 and R52 is meant to vary depending on the thermistor resistance (it's a potential divider). In my case this still happened (regardless of the higher resistance from the damaged R52), however this junction voltage was held low by the damaged D19. Consequently, the non-inverting input to U8 was maintained close to 0V.

As the temperature of the thermistor increases, the resistance of the thermistor decreases (from the service manual) and the voltage output of the potential divider provided between R317 and R52 is meant to decrease. Hence, a decreasing voltage input into U8 indicates, to the MCU, an increasing temperature. In my case this voltage was low, very low, being held in place by D19, which indicates a very high temperature to the MCU which I reporting in my OP:

...the SCI output the tenths-deg-C value is 797


Without a battery connected, the output of the junction of R317 and R52 is meant to be held at 5V since no potential divider forms and R317 holds the voltage high. As a result the input to U8 is also at 5V. I guess the corresponding output indicates to the MCU that no battery is present.

In my case, with a damaged D19, even without a battery present, D19 holds the voltage input into U8 close to 0V and consequently the MCU thinks a battery is present (since it expects a high output to indicate a lack of battery). The low voltage also indicates an enormously high temperature. With external power applied, the MCU enters into charging mode and waits for the battery to cool down, which of course never happens, so no current flows into the battery as the charging fets are not active.

Does that make sense? A schematic would help, I'll see if I can make one.
blevine
 
Posts: 12
Joined: September 12th, 2012, 4:57 am

Re: Roomba 500 R52

Postby mfortuna » September 21st, 2012, 6:54 am

A question for Gordon, is D19 protecting the op amp or the MCU? A single rail opamp output can't go negative. I would guess they are using the opamp to scale the thermistor voltage into a 0-5V range for the MCU ADC. In this case D19 protects the opamp input, I think.
Mike
Reds x 3, Dirt Dog, Disco (now a parts bot), Create, Scooba 350, and Security Dawg
Evolution Mint
Neato XV-11
Shark Ion 750
User avatar
mfortuna
Robot Master
 
Posts: 5844
Joined: February 5th, 2006, 10:35 am
Location: NH

Re: Roomba 500 R52

Postby blevine » September 21st, 2012, 7:09 am

I checked the datasheet for the LM324 opamp (there are 4 in U8) and it mentions the input voltage into either input cannot go below -0.3V.

I also measured the supply voltage for U8 and got 7V and 0V. Does this mean the input to the MCU could go as high as 7V? I guess you'd expect a couple of volts drop in the LM324 itself.
blevine
 
Posts: 12
Joined: September 12th, 2012, 4:57 am

Re: Roomba 500 R52

Postby blevine » September 21st, 2012, 9:14 am

Just some rough schematics based on Gordon's values and ASCII drawing.

Temp_sense_bat_con.jpg
Roomba 530 Temp sense circuit with battery connected


With the battery connected a potential divider forms from R317, R52, R319 and the thermistor.
Some very rough values for output voltages are shown in purple. As the temperature increases the voltage reduces. These are based on resistance values for the thermistor in the service manual and Gordon's values.

These voltages are fed into the non-inverting input of an opamp in U8. Any components on the inverting input have not yet been investigated.

In my case R52 was around 1.6k so any temperature value would be slightly off. However, the bigger problem was a short circuiting D19 which caused a value close to 0V at the junction of R317 and R52. The MCU interpreted this as a high temperature of around 80 degrees C.

Temp_sense_bat_discon.jpg
Roomba 530 Temp sense circuit with battery disconnected


With the battery disconnected the potential divider is eliminated and hence R317 holds the voltage at 5V. This is fed into the non-inverting input of an opamp in U8. This must indicate to the MCU that no battery is connected.

In my case the faulty D19 still holds the voltage close to 0V at the junction of R317 and R52. Hence the MCU still detects a battery and enters charging mode if external power is supplied. Charging does not occur due to the perceived high temperature.
blevine
 
Posts: 12
Joined: September 12th, 2012, 4:57 am

Re: Roomba 500 R52

Postby mfortuna » September 21st, 2012, 11:26 am

blevine wrote:I checked the datasheet for the LM324 opamp (there are 4 in U8) and it mentions the input voltage into either input cannot go below -0.3V.

I also measured the supply voltage for U8 and got 7V and 0V. Does this mean the input to the MCU could go as high as 7V? I guess you'd expect a couple of volts drop in the LM324 itself.


It depends what the gain is set to. If it is unity gain then there will be some drop through the opamp but at low currents like a MCU input I would expect a transistor drop of < 1V.
Mike
Reds x 3, Dirt Dog, Disco (now a parts bot), Create, Scooba 350, and Security Dawg
Evolution Mint
Neato XV-11
Shark Ion 750
User avatar
mfortuna
Robot Master
 
Posts: 5844
Joined: February 5th, 2006, 10:35 am
Location: NH

Re: Roomba 500 R52

Postby Gordon » September 21st, 2012, 12:54 pm

Wow! So much activity while I am trying to awaken. I must leave to do morning chores in 15 mins so will be brief as I try to get some items sorted out:

Thermistor: All of iRobot's thermistors are negative temperature coefficient, resistance lowers as temperature increases. blevine's SPICE schematic, shown next, shows that signal behavior:
schematDWNbyBlevine.jpg
<reduced image res on 1209211530> via the listed three sample temperatures.

Signal at R317 / R52 tap: When no thermistor loads this point, the signal is ~5V and should tell the MCU the battery is too cold to charge. If a shorted D19 had been connected to that tap the MCU would again not start charging due to a too hot battery.

U8 Op Amp: U8-9(-IN3) is connected to U8-8(Out3), making the amplifier a VOLTAGE FOLLOWER. There should be no more difference in input vs. output voltage than about 5mV (or whatever the differential input voltage error of the LM324 amounts to). {Text has been trimmed to correct wrong referral.}

D19's intended service: Apparently both ICs (U8[A3] and U28-84) benefit by clipping any below zero-volts (say, due to a noise pulse) swing of the signal line.

-----------------------------------------------
Now that blevine has drawn my attention to the signal voltage being on the order of two to four volts for our derived circuit, a point I can't argue against in the as-drawn schematic. Therefore, I tend to think some part of the circuit may be missing! I say that because I have been measuring signal voltages for a range of simulated thermistor resistance, e.g.: Node-MCUport. */ Signals = 2.73@10k, 2.46@8k, 2.11@6k, 1.64@4k, 1.00@2k and 0.57@1k /*, on a loose 560 main_PCA, and we can see the signal occupies a lower voltage range. More work is required to sort this difference.
Last edited by Gordon on September 21st, 2012, 6:31 pm, edited 1 time in total.
Gordon
Robot Master
 
Posts: 4304
Joined: April 6th, 2005, 2:02 am
Location: Santa Ynez, CA USA

Next

Return to iRobot Roomba and Scooba Chat

Who is online

Users browsing this forum: dboston55, fufu, jdong, leonicholson, Metalworker, someoneclever, vic7767 and 414 guests