Looking for a Home Base Schematic

[Tigertron]

My home base quit charging but the disco charges fine through the rear connection. I figured it must be the MOSFET and I replaced it only to find it still does not charge. Deeper troubleshooting shows that a 3.3k resistor across the terminals does not put it into docked mode although the 4.5vdc is there. I can force the MOSFET on and the battery does charge. In an attempt to reverse engineer the docking circuit I found two chip resistors R39 and R40 that seem to be marked 430 and 410. Which means they should be 43 ohms and 41 ohms but in fact measure (in circuit) over 180k. A schematic of this circuit would really save my eyes. Has anyone got one?
Thanks a million.

[Gordon]

...A schematic of this circuit would really save my eyes. Has anyone got one?...

Great work so far. Looks like you've had to go farther than anyone else!

There's been no HB-schematic published, AFAIK. Making that schematic has been on my to-do list for years. It obviously has low priority, hence, I doubt I'll ever get around to it!

If I can cut loose some time in the next day or two, I'll try to repeat your in-ckt resistance measurements on the two resistors.

[Tigertron]

That would be great. I saw the detailed schematic you did of the Disco. That is some fine work. I was hoping someone had done the same or even a scratch pad drawing of the HB. No luck it seems.

Looking at the PCB I see that those resistors are in series between ground and the + output to the battery. The junction of the two resistors heads off to a input (pin 7) of the LM339 (quad opamp). I see feedback resistors and it looks like its a comparator which could be part of the docking sensor I suspect but with out a schematic its real hard to tell.

Perhaps, if you would be so kind, you could measure a voltage there as well. It is labeled T28. Also pin 6 and pin 1 voltages of the LM339 would be helpful.

If you can't, don't worry. I may have to sit down with an ohm meter, a magnifying glass and a pad of paper to see if I can figure this out.

Thanks again.

[Gordon]

This is all I have time for today, until late night.

I found two chip resistors R39 and R40 that seem to be marked 430 and 410.

Yes, you will have to use higher power optical magnification! R39 is marked "43D", and R40 "41C". They are special value, or at least, 1% tolerance devices.

My resistance checks, in-circuit, show 146.7k across R39, 26.0k across R40, and 133.6k ! across both (contacting their left ends).

... The junction of the two resistors heads off to a input (pin 7) of the LM339 (quad opamp). I see feedback resistors and it looks like its a comparator ..

Yes, the 339 is a quad op-amp designed for comparator service. iRbt uses them all over the place. When an ordinary op-amp is required, its the LM324 that you'll see.

... you could measure a voltage {on} T28. Also pin 6 and pin 1 voltages of the LM339 would be helpful. ...

I'll work on those measurements later.

[Tigertron]

Ahh. Very good! Yes I need better magnification. Getting the probes on a little better they ohm out to those values. It didn't seem right for them to be only 80 ohms total. I couldn't imagine them even going bad at that. This weekend I plan to tear into it deeper. Perhaps I can come up with at least a line drawing to share. I have since verified that the source 5 vdc is good.

Thanks for your input Gordon, it is much appreciated.

[Gordon]

Working with a 2004 vintage Home Base having a PWB asm with the following identifiers:

Assembly P/N: 74336-003-0021(09-09)
DOC_BOARD_6PCB
09 JULY 2004
SCH V 8.3,

the following measurements were made at selected U4 (LM339) pins, and one test-point (#28), which is common to the R39, R40, and C17 node.

Conditions were:
a) Powered via Fast Charger SMPS.
b) HB's Power LED lit.
c) 'Docked' LED dark.
d) Roomba's charging interface: normally open-circuit.

All values are wrt "T_GND", test-pad at lower left-corner of the PCB:

V_T28 = 0.38 V

V_U4-1 = 4.96 V (OUTPUT 2)

V_U4-3 = 5.03 V (VCC)

V_U4-6 = 2.52 V (INPUT 2-)

V_U4-7 = 2.73 V (INPUT 2+)

V_U4-12 ~ 0 (GND) Fluctuating (probably) due to IRED-drivers' pulsing.

U4 pins 3 & 12 were measured to confirm the other pin numbers were true.

[Tigertron]

Thank you Gordon. At least now I have some reference points to begin from. I am going to work on it later today or over the weekend. Hopefully I can figure something out.

[Gordon]

Tigertron, I took a cut at casually extracting part of the circuit that up-switches charging voltage. The first draft was full of errors. I will provide a better crafted draft as an attachment to a subsequent post in this thread. Meanwhile, you may be able to use a batch of voltage samples that I obtained.



Following each test-point's name / descriptor, its coordinate location is approximately recorded as x & y offsets from the lower left corner of the PCB. Units are millimeters.

[Tigertron]

That is great. Thanks again. I have not had the chance to troubleshoot yet. I have to finish up a painting project first. My wife reminded me we have guests coming for Thanksgiving and since the roomba is charging from the back port I will have too look at it a little bit later. Troubleshooting is my fun so I'm sure I'll poke at at least a little bit tonight.

Hope your computer feels better.

[klapskalli]

Hey Gordon,
did you get to create a schematic of the home base? I've bought a used DISCO and it's home base isn't working anymore. With the voltages in this thread i was able to find out that the regulator (K78L05A) right to the point "T_VCC" was broke and replaced it.
We replaced the capacitors (4.7muF & 100muF) around it, too. There definitely are 5.02v on the "T_VCC" point. But the home base still doesn't work right:
- Power LED lights up, now
- docking LED is always on, even when not docked
- there is no 4.7v voltage on the charging contacts
- and no 22v when docked.

I'm not sure, if we set the capacitors. I think it is possible, that we mixed them up (left <-> right). Now, the 100muF capacitor is closest to the edge of the board, the 4.7muF capacitor closer to the middle of the board. Is that right?

Might this be a problem? If so, can someone tell me, which capacitor goes where? An image of the original board, indicating which capacitor sits where, might be helpful. If the capacitors are correct, I know I have to look elsewhere for the error.

Thanks in advance from germany. Tim

[Gordon]

...did you get to create a schematic of the home base?

I have pencil drafts of the HB's circuit sections, but have no plans to make digital images of each section. I think the work involved is not warranted. During the last few days that I was able to work on the HB-project I formulated the idea of preparing a quasi block/schematic diagram of the entire HB, which I would then post in this thread. The pencil-sketch of that diagram requires more work, but if you need a copy I could try scanning it, then attaching it to a PM for you. I expect some weeks will pass before I will be allocated time to make a clean CAD image of the diagram.

I've bought a used DISCO and it's home base isn't working ... i was able to find out that the regulator (K78L05A) right to the point "T_VCC" was broke and replaced it. ...We replaced the capacitors (4.7muF & 100muF) ... too.

I'm not sure, if ... we mixed them up (left <-> right). Now, the 100uF capacitor is closest to the edge of the board, the 4.7uF capacitor closer to the middle of the board. Is that right?

Looks like they have been mixed! Here are my notes for those two:

C1[1,1]+(13,27){4.7uF,50V}==>U1-OUT==>+5VREG distribution,
...
C5[1,1]+(24,24){100uF,35V}==>U1-IN==>U2-S, etc.,
for which the numbers between (..,..) are the x,y coordinates, millimeters, as measured from the lower left (viewed from component-side) corner of the PCB to middle of a component's footprint. Here's a pic of that portion of the PWB-asm.:

(the cruciform annotations demark GND, all are BLK except two aqua overlaying dark backgrounds)

There definitely are 5.02v on the "T_VCC" point. But the home base still doesn't work right:
- Power LED lights up, now
- docking LED is always on, even when not docked
- there is no 4.7v voltage on the charging contacts
- and no 22v when docked. ...

Tim, I expect those three features will all work as soon as you discover/correct the remaining fault(s).

One big thing to worry about is: What took out that regulator? IOW, was it something big enough that many other components were killed at the same time? I will insert below a draft of all the test-point signals/dc-voltages which I have logged. Same for pins on the quad-comparator, U4, and the iRbt-chip-on-board IC, U3.

You might want to concentrate on checking values on test-pads T24 through T29, since you know T30 is not right. T24 & T27 are voltage-divider points, and should show half of your +5Vdc supply. T25 & T28 are also dc-volts divisions, but not 5V/2 (see the inserted table). T26, 28, & 29 are semi-dynamic during docking, and may require a scope to assess properly.

V_T30 will be 4.3Vdc, NOM, with a powered HB in the un-docked state. That voltage is developed by impressing the +5Vdc supply across R16(5400) + D10(fwd biased, & K==>T30) + R39(144k,ic) + R40(26k,ic) to GND. Lower-case "ic" = 'in-circuit'.

Look in the following table to see what the dynamic nodes do before docking, during docking, and after docking:

Code: Select all

T1;U3-8; (same as 'JUMP_1")
T1(25,35) V_T1 = 0 to 5Vpk, 38kHz SQ-wave (noisy & w/over-shoot).
          IoDS (Independent of Docked State).

T2;Q2c;
T2(27,45) V_T2 = 0 to 22Vpk, 38kHz SQ-wave, (wrinkled crest) IoDS.

T3;U3-16;==>R4:R4==>Q1b
T3(23,33) V_T3 = 0 to 5Vpk ~SQ-wave (75% HI,25% LO), with 4.5 to
5Vpk, 38kHz ~SQ-wave, 65% HI, 35% LO. If state=doc'd, V_T3 = 0V.
See T1vsT3signals.jpg, or .apc. T3 shows raw Lft-bcn's signal.
Compare to radiant waveform.

T4;Q1e;
T4(13,31) V_T4 = 0 to 4.2Vpk ~SQ-wave (75% HI,25% LO, with 0.3V to
4.2Vpk, 38kHz ~SQ-wave 65% HI, 35% LO. If state=doc'd, V_T4 = 0V.
See T4_Ndocd.jpg, or .apc, & T4_Ndkd_pulsePkt.jpg. T4 provides a
measure of Lft-bcn's drive-current:
<I_lft_bcn> ~ (.75)*(.5)*(Vpk_T4 - Vmin_T4) / 100.
Lft-bcn's IRED-current sensing (100 ohm emitter res.).
Compare T4_WF to radiant waveform.

T5;U3-15;==>R18:R18==>Q7b
T5(62,40) V_T5 = 0 to 5Vpk PWM wave, with 4.5 to 5Vpk, 38kHz ~SQ-
wave. 8-bit words occupy NOM 40ms frames. A 'one-bit' is described
by 3ms @ 5V + 1ms @ 0V, and a zero-bit by: 1ms @ 0V + 3ms @ 5V.
The RIGHT-beacon's source (U3-15) signal is 11110100 = $F4 = 244
decimal. If state = doc'd, V_T5 = 0V. See T5_Ndocd.jpg, or .apc.
Compare to radiant waveform.

T6;Q7e;
T6(58,44) V_T6 = 0 to 5Vpk PCM wave, with 0.5 to 5Vpk, 38kHz ~SQ-
wave. As with T5 the RIGHT-beacon's PWM code is 11110100 = $F4 =
244 decimal. If state = doc'd, V_T5 = 0V. See T6_Ndocd.jpg,
or .apc. Compare to radiant waveform. Right-bcn's IRED-current
sensing (100 ohm emitter res.).

T7;U3-14; ==>R14:R14==>Q5b
T7(52,39) V_T7 = 0 to 5Vpk PWM wave, with 4.5 to 5Vpk, 38kHz ~SQ-
wave. 8-bit words occupy NOM 40ms frames. A 'one-bit' is described
by 3ms @ 5V + 1ms @ 0V, and a zero-bit by: 1ms @ 0V + 3ms @ 5V.
The Frc-Fld-beacon's source (U3-14) signal is 11110010 = $F2 = 242
decimal. If state = doc'd, V_T7 = 0V. See T7_Ndocd.jpg, or .apc.
Compare to radiant waveform.

T8;Q5e;
T8(39,39) V_T8 = 0 to 4Vpk PCM wave (clean). As with T7 the Frc-
Fld-beacon's source (U3-14) signal is 11110010 = $F2 = 242
decimal. If state = doc'd, V_T5 = 0V. See T8_Ndocd.jpg, or .apc.
Compare to radiant waveform. Frc-Fld-bcn's IRED-current sensing
(47 ohm emitter res.).

T9;U3-13; ==>C14[1,2]
T9(68,41) V_T9 = 0Vdc Ndoc'd, 4.5Vdc doc'd, for Docked-LED drvr.

T10;Q9c; Docked-LED driver
T10(71,46) V_T10 = 3.5Vdc Ndoc'd, ~0Vdc doc'd & Docked-LED ON.

T11;F1(OUT);
T11(15,14)  V_T11 = 22.0Vdc, NOM, input from Fast Charger SMPS.

T12;U1(IN+);
T12(19,26)  V_T12 = V_78L05(IN+) = 20.8Vdc (w/~0.3V ripple).

T13;U3-5;
T13(63,13) V_T13 = 0 to 5Vpk, 32Hz SQ-wave, IoDS.

T14;D5K;
T14(59,19) 0.86 < V_T14 < 3.2V odd wave, at 32Hz. An IoDS periodic
pulse-train is seen at T14. Its best to view the scope trace (see
T14_Ndoc.jpg) because precisely describing the wave needs many
words. One may try by visualizing pulses formed by slightly
differentiating a 32hZ square-wave. Its rise is no longer instant,
but takes about 7% of the overall pulse-width to reach the 3.2V
peak. Rolling over the peak, there is a -0.6V droop (spanning
~half the pulse-width), after which voltage drops in the manner of
an R*C discharge curve to zero, completing one pulse in the wave-
train.

T15;Q4e & Q6c;
T15(43,28)  V_T15 = 0.015Vdc {~0V} Ndoc'd. When state=doc'd, the
signal changes to an ~saw-tooth waveform (SEE T15_docd.jpg)
exhibiting ~0.08VPP, 32Hz.

T16;U3-4;
T16(62,15) V_T16 = 0V, Ndoc'd, but when doc'd V_T16 goes HI to 5V,
and narrow timing notches fall to zero at 5Hz rate. See
T16_docd.jpg. Timing-notches are 2ms wide.

T17;Q4c;
T17(37,23) V_T17 = 21.6Vdc, Ndoc'd, but when state = doc'd V_T17
goes LO to zero, and with narrow timing pips rising to 5Vpk at 5Hz
rate. Timing-pips are ~2ms wide.

T18;U3-3;
T18(60,11) V_T18 = 0 to 5VPP, 32Hz SQ-wave, IoDS.

T19;U4A2-6{IN-};
T19(74,38) V_T19 = 0.3VPP, 32Hz sawtooth, IoDS. See T19_Ndocd.jpg.

T20;U4A2-7{IN+};
T20(89,24) V_T20 = 2.73Vdc, IoDS. This is a fixed reference
voltage for the comparator's non-inverting input.

T21;U4A2-1{OUT};
T21(85,35) V_T21 = 4.96V, IoDS. U4A2's function remains unknown.

T22;Q8b;
T22(65,37) V_T22 = 0.01V, IoDS. Q8's function remains unknown.

T23;D9K;
T23(62,33) V_T23 = 0.01V, IoDS. This function remains unknown. An
input should reside on this node, but there is none!

T24;U4A4-11{IN+};
T24(97,29) V_T24 = 2.93Vdc, IoDS. This is a fixed reference
voltage for the comparator's non-inverting input.

T25;U4A3-8{IN-};
T25(96,24) V_T25 = 2.00Vdc, IoDS. This is a fixed reference
voltage for the comparator's inverting input.

T26;U3-12;
T26(100,15) V_T26 = 0Vdc while state = Ndoc'd, however, during
docking a single, ~19ms wide, 4.8V pulse occurs.

T27;U4A1-5{IN+};
T27(88,30) V_T27 = 2.52Vdc, IoDS. This is a fixed reference
voltage for the comparator's non-inverting input.

T28;
T28;U4A1-4{IN-}; AND T30
ChA = T28(95,4) = U4-4 = U4A1-4{IN-},
ChB = T30(65,5) = D7K, at moment of simulated docking.
Docking was induced by switching 4.7k ohms across HB-output
terminals at "J2".
T28(cont.)
V_T28 is the fractional sample (~8.5%) of V_T30 that is input to
the inverting input (pin-4) of comparator U4A1, (section-1). That
input signal is compared to ~1/2 of +5VREG, which biases non-
inverting input pin-5 of U4.
T28(cont.)
Scope measurements reveal the following actions as a docking event
takes place. Note that these data derive from a simulated docking
in which no charging current flows, only a charging-level voltage
is switched ON.
T28(cont.)
TP Sig-  Pre-doc'g   Doc'g  Post-doc'g     Comments / Notes
Name     (volts)   (volts)  (volts)
V_T28      0.37      0.2       1.85   dt_Dock'g ~ 30ms wide,[(1)]
V_T30      4.37      2.0      21.5    dt_Dock'g ~ 30ms wide,[(1)]
T28(cont.)
----------------
T28 & T30 NOTES:
[(1)] Both Post-doc'g signals are seen to be merged with periodic
switch-off pulses of ~1ms width and occurring every 200ms. These
pulses are injected via U2's gate-driver.
----------------

T29;U3-11;
T29(98,10) V_T29 = 5Vdc, IoDS, but it should go LO when U4A1-OUT
goes LO. The function of U4A1 remains unknown.

T30;D7K;
T30(65,5)    See above T28 & T30 combined data.

T_GND;
T_GND(10,2)  0.00V NOM reference point, only.

T_VCC;
T_VCC(14,22) +5.00V NOM reference point, only.


The following non-TP device-pins' data may also be of interest.
U3(74,14);
U3-1 V_U3_1 = 0Vdc, IoDS.

U3(74,14);
U3-2 V_U3_2 = 4.96Vdc, IoDS.

U3(74,14);
U3-6 V_U3_6 = 0Vdc, IoDS. This is COB GND.

U3(74,14);
U3-7 V_U3_7 = 5Vdc, IoDS. This may be COB VCC.

U3(74,14);
U3-9  V_U3_9 = 5Vdc, IoDS. This is a soft +5V bias into U3.

U3(74,14);
U3-10 V_U3_10 = 5Vdc, IoDS. This is a soft +5V bias into U3.

U4(93,30);
U4A1-2 V_U4_2 = 5Vdc, Iods.

U4(93,30);
U4-3   This is U4's VCC = +5Vdc.

U4(93,30);
U4A3-9 V_U4_9 = 4.78Vdc when state = Ndocd. If state = docd, then
~1ms wide timing notches fall from the 5V level to 2.7 volts every
200ms.

U4(93,30);
U4A4-10 V_U4_10 = V_U4_9 since pins are on the same node.

U4(93,30);
U4-12 This is U4's GND.

U4(93,30);
U4A4-13 V_U4_13 = 0Vdc when state = Ndocd. If state = docd, then
~1ms wide timing notches rise from the zero level up to 4.8V every
200ms.

U4(93,30);
U4A3-14 V_U4_14 = V_U4_13 since pins are on the same node.

U2(32,28);
U2[1,1]G V_U2G = 21.6V when state = Ndoc'd. If state = doc'd,
then V_U2G pulls down to 2V for 199ms then rises to 21.6V for
1ms, and repeating that sequence every 200ms. The FET conducts
while the gate voltage is LO.


Working w/o a schematic makes it much tougher! If necessary, I can supply my track-tracing notes to you (from which you could draw your own schematics), or try scanning pencil sketches. I trust you are messing with this HB for the fun of it, rather than being pressed to get it working! However, if the latter, you might consider asking ProTechRobotics-Chris what he would charge to separate an HB's PWB-asm from its housing (& counter-weights), then ship just the board-assembly to you.

[klapskalli]

Hey Gordon,
thank you VERY much! Now i know where to look at next/again. And your table will definitely help me to get it working again, i think.

I trust you are messing with this HB for the fun of it, rather than being pressed to get it working!

Well, the home base isn't working and the DISCO's rear jack neither (that'll be the next thing on my list...) So, i can only charge the roomba, when i short-circuit the plug of the APS with the contact of the home base. But i was lucky on ebay and got a broken DISCO with a working home base / virtual wall / APS / remote, yesterday. It was double the price of a new home base, but now i'll have enough spare parts for the roomba and can play around with the home base electronics.

Who knows, maybe there is even a chance to repair that roomba... There definitely seems to be enough knowledge and helpfulness on this board!

Again, thanks a lot from germany. Tim

[Yorlik]

Hi, has anyone ever made a schematic of the home base? I have 3 broken ones in various states of misrepair

Thank you.