Spacedoggys WB 2010 Build Diary

[spacedoggy]

Hi folks, This is a new thread for my WB 2010 build I'm doing in parallel of the original WB in the other thread.
since I had a bit of trouble with the 5V rail on that one, I thought I could try building another one for trouble shooting reference.

so, built the power supply stage, and everything worked perfectly, very little variance and consistent power supply from the battery board to the rails.

went on to the controller stage, atmega took a bit of tweaking to get the .hex uploaded, I can only seem to be able to get AVR studio 4 as the atmel site sucks.
anyway, that defaulted the fuse divide internal clock by 8, the result being the serial input rate for the terminal was also divided by 8, and all the tests were really slow.

i fixed that and the blinky led was normal looking, but i notice the PWM tests and other input phase tests were (approx 8x) slower than on the original WB. hope that's normal with the new code...?

also the potentiometer test is similar to the vid i posted on the WB 1 diary? sawtooth decreases to 0V. being the case on both units i decided to proceed.

-I attached all remaining components on the underside
-and the VCO switch topside.

I probed the VCO inputs: G and H on the WB 2010 menu

and notice the power falling from 5V to 3V on VCO 1input ... and from 5v to 0V on VCO 2 input...

I made a vid of me probing the stable rails 3.5, 5, 12, 21
and then running G and probing VCO1
and then running H and probing VCO2

http://www.youtube.com/watch?v=U307psD7uTc

is this normal? again I'm reluctant to proceed adding my expensive VCOs until I'm sure this part is ok.

[Lcstyle]

The Vcc for both VCO's is controlled by the mic2514 which are switches turned on by the 168.
I believe the test code g and h cycles the power every 2 seconds, so you need to probe with an oscilloscope.

Still even with a volt meter, the power displayed should alternate between +Vdd (mic2514 mosfet) and 0 every 2 seconds.

Probe IC11 pin 3 to ground and IC10 pin 3 to ground. Should always be +5V


/**
* Test VCO1 power switch
*
* Cycle VCO1 power every 2s.
*
*/
void test_powerswitch1(void) {
POWERCTL2_PORT &= ~_BV(POWERCTL2);
POWERCTL1_PORT |= _BV(POWERCTL1);
delay_ms(2000);

POWERCTL1_PORT &= ~_BV(POWERCTL1);
delay_ms(2000);
}

[Mictronics]

@ spacedoggy

Did you set the fuses like shown on the wiki page?

[Lcstyle]

@ spacedoggy

Did you set the fuses like shown on the wiki page?

I used
http://www.engbedded.com/fusecalc/

and got E2 D5 F9
I used boot flash section size 1024 instead of the 128 default.

You definitely don't want div/8

[spacedoggy]

@ spacedoggy

Did you set the fuses like shown on the wiki page?

yeah, but the layout in AVR studio 4 is different than the screen shot.

I used
http://www.engbedded.com/fusecalc/

and got E2 D5 F9
I used boot flash section size 1024 instead of the 128 default.

You definitely don't want div/8

I'll check this tomorrow evening. nice way to be sure.
avr dude doesnt play well with my programmer i need to install winlibusb driver which kills avr version
avr studio 5 can be downloaded, but it mails me a bum link after i register.

thanks for feedback.

[Lcstyle]

@ spacedoggy

Did you set the fuses like shown on the wiki page?

Umm, dude just use avrdude

http://www.ladyada.net/learn/avr/avrdude.html

and

http://www.ladyada.net/learn/avr/fuses.html

[spacedoggy]

ok, decided to load up avr dude again and tried this

avrdude -c avrispmkii -p atmega168 -P usb -U lfuse:w:0xe2:m -U hfuse:w:0xd5:m -U efuse:w:0xf9:m

ran through quick, but here's the important part...
(wouldnt allow redirect of output to a file and XP dont like scrolling up, lame)


(if i hit Y the orange light on the programmer goes on and nothing happens)

meanwhile, my usb to serial controller is acting the bollocks on 2 battery boards. and for some reason when I ask it to power down using the Z command it starts again straight away without me pressing the power button.

I feel like throwing this thing out the window! Grrrrrr

gonna get some sleep and come back with a fresh head.

any help greatly appreciated as always.

[Lcstyle]

Sel

ok, decided to load up avr dude again and tried this

avrdude -c avrispmkii -p atmega168 -P usb -U lfuse:w:0xe2:m -U hfuse:w:0xd5:m -U efuse:w:0xf9:m

ran through quick, but here's the important part...
(wouldnt allow redirect of output to a file and XP dont like scrolling up, lame)


(if i hit Y the orange light on the programmer goes on and nothing happens)

Please note that for these values:
* Note that some numerical values refer to fuses containing undefined bits (set to '1' here). Depending on the target device these fuse bits will be read either as '0' or '1'. Verification errors will occur if the values are read back with undefined bits set to '0'. Everything is fine if the values read from the device are either the same as programmed, or the following values (undefined set to '0'): Extended: 0x01.
from: http://www.engbedded.com/fusecalc/

"The more shortcuts you take the longer your journey will be."

[spacedoggy]

ok, avrdude used to rewrite the fuses, blinky light test is faster than normal, but I notice that the thing does not turn off now...

I apply power --> off, (ok)
press power button --> on
z in terminal --> reboot/on

everything still works so I'll keep going, probably bad transistor. I'll trouble shoot after i get the blasted thing working.

___________________________________

I installed the VCOs and fine pitch IC to the board, all that remain now are the tiny resistors and caps. being hard of sight this work is verging towards the microscopic. I fixed the trace above VCO2 and am carefully avoiding the minefield of shorts to ground. continuity testing all connections is difficult on low value resistors, although the multimeter beeps indicating a short to ground, the measure of resistance is is displayed.

I'm going to finish the PLL and amplification sections, then run all the final tests.

_____________________________________

for R20 and R21 I have 30K resistors installed as in the wiki, the Lcstyle info post suggests i need to change these values to resistors I don't have at the moment, will it still work with the default 30K resistors? perhaps we can update the wiki with the info from his post.

______________________________________

some comments on the WB2010 design:

The buttons. I find it strange about the footprint of these buttons. there are only 2 tiny traces holding these components on and I'm afraid they will break of the button is pressed hard enough. I was thinking of breaking out these buttons using wires to an external case.

Cap on the battery board between the usb conector and serial IC (C7 I think) is a pain in the ass to get at and remove, I had the situation where it was over heating and found it very difficult to get at.
______________________________________

Going forward: testing, since i done have a spectrum analyser, testing after VCO stage might be difficult. I will run the PLL tests from the original tutorial as best i can, are there any other methods i can use to test this badboy? set it to block an FM frequency or something I can detect with household equipment?

better still, can i build a spectrum analyser on the cheap. I found a kit to build a function generator that saved me buying expensive bench equipment. would be nice to assemble other diagnostic tools myself cheaply instead of paying for tools for 1 or 2 projects.

Pics comming soon.

SD.

[spacedoggy]

yo dudes...

I'm happy out, I completed the PLL and amplification stages, WB 2010 hardware is built
I still have post VCO testing to do and to iron out the shut-down transistor bug before I can load the final firmware,
but I'm happy to have placed all the components and have power running without melting or smoke.



as promised, hi res pics on the way when i borrow the DSLR again.

for now sleep.



SD.

[Lcstyle]

Do the VCO's Lock on the VCO1 and VCO2 Tune Test?

[spacedoggy]

I'll run tests later after work. The tests i can that is. I don't have a spectrum analyser, too expensive for me. but will do what i can. What kind of lock up am i looking for? anything I can check here that might help you is the least I can do.

also for R20 and R21 i still dont have the values you specified for 1300 and 2700... I guess I'll try a workaround there.

PM me your chat details and perhaps we can talk later?

SD.

[Lcstyle]

Use the formula I posted to determine what the resistor values should be for the VCO VTUNE.

It will work with the existing resistors, but you may not be able to tune the full range of the VCO.

To see your VCO lock, put your scope or DMM on the vtune pin of the VCO. Using the test software, attempt to use the tune VCO function. You should see a slowly rising Voltage all the way from 0 to 25V (on the 2700 with the correct resistor values) look at the datasheet for the VCO, once the PLL finds a lock the VTUNE Voltage should stabilize around a certain range. For the 2700 I think I remember this was from 8-9 V for the Test Software Freq.

[spacedoggy]

OK local electronics shop doesnt have the values i need, so sticking with the 30K resistors for now...

found 2.5V chip was bad so i changed it out, seems it was shorting somewhere because after i replaced it my power switch bug was fixed.

crystal shows sine wave, PLL pin 10 shows blinking square wave, cool

next i ran:
p - PLL test 2 - RF stage
q - PLL test 3 - IF stage

which returned the following lines...

PLL for RF freq 2000MHz & prescaler 8: B=250 A=0
PLL for RF freq 1200MHz & prescaler 8: B=150 A=0

...not sure what that means.

next i run
l - VCO sweep test

it both sweeps the VCOs vtunes from 0V to 13.2V, then back to 0V and repeats, it does not stop or settle on any value.
as in ladyada's tutorial, I probed the 10K resistors below the PLL R37 and R38 during the sweep. the value is a constant 2.3 V

TBH I'm not sure how to progress here. the wiki says that the wave bubble menu is self explanatory, yet I'm not sure about validating
this final tuning test, or how to load a full firmware to test it by jamming something.

as always, your thoughts and comments appreciated.

[spacedoggy]

ok, I got by power stabilised, and resumed work on this.
now using an external bench power supply and a sparkfun FTDI connection to the PC


I reset the fuses, and loaded the test firmware and ran voltege, PWM 555 and digi pot tests, all ok

next i replaced:
-R21 with a 47K resistor (0V - 18.81V to the 1300 VCO*)
-R20 with a string of 33K + 22K + 10K = 65K (0V - 24.75V to the 2700 VCO*)

next I ran the VCO sweep test.

1300 does not lock
2700 locks at 20.4V

2 questions:
-What do i need to check to get the 1300 to lock?
-the 2700 locking at 20.4V seems a little high, is this ok?


_____________________________________________
*using formula 3.3*((X/10)+1)=Y
where X is Kohm
and Y Voltage limit of the sweep test