Nikkai 5.8GHz Video Sender

[wd40addict]

I know what you're thinking - this isn't vintage! Well no, it isn't but, I use it to feed the Aurora that feeds my TV22 and TMB272 so I trust it will be of interest. It also shows it's still possible to fix modern equipment that on first sight seems hopeless.

For sometime I've used a 2.4GHz video sender to send TV pictures from the Tivo in the living room to my den. Of late this hasn't been that satisfactory as there seems to be so much other interference in the 2.4GHz ISM band. It's only 100MHz wide and is now a free-for-all with WiFi, Bluetooth, security cameras etc etc.

I have been idly wondering whether the 5GHz band would be better as it's wider, used by less equipment and has shorter range (so less signals from the neighbours). These days people want to send HDMI signals so the older analogue types are being sold on. Nikkai was the in-house brand at Maplin presumably named to sound Japanese. I picked up a 5.8GHz video sender (Tx and Rx units), Rx shown in the picture, but Tx looks identical. It was claimed to be working, but on plugging into a modern TV the sound had a loud squeal superimposed on it and the picture was a feint rolling mess :-(

Rather disheartened I started digging in to see what I could find. The units are held together by one visible screw and two hidden under the front feet. Once inside I found a small screened module mounted on a through hole style 'motherboard'. Photo shows the inside of the Rx, but Tx is similar. First thing I noticed was the slight browning of the PCB around the 7805 regulator. These units seemed to have been powered up 24/7.

The pins to the Module are silkscreened with labels, poking around with a scope showed a large ~1.5kHz signal everywhere, but greatest on the 'bypass' pin. This appeared (and was later confirmed to be) where a 470uF cap was connected to the output of a regulator within the Tx module. The 470uF wasn't in great shape, but substituting a good part made no difference, the squealing could only be cured by using a much larger cap and even then the voltage was 3.2V which didn't seem right.

Unclipping the tin lid of the module revealed the circuit to be based round an AWI5800 IC. Some Googling found a datasheet, which included a typical application circuit, this was suspiciously similar to my unit...

My speculation was that the capacitor had been dried out by the heat and this had somehow damaged the module's surface mount regulator. As I had nothing to lose I tried feeding in 3V3 from an external power supply - result perfect audio!!! The damaged regulator was an obscure part, but I realised that I could remove the 5V feed from the module and just feed an externally regulated 3V3 into the 'bypass' pin. The existing regulator would just remain a passenger doing nothing.

So I now had good sound with no interference from any neighbours. The PCB was littered with 470uF caps and any that were duff were replaced. Note how some are positioned right next to the hot regulator...

So I now had good sound, but a crappy picture. Feeding a sig gen into the Tx video input and observing the output from the Rx on a scope didn't show any obvious issue apart from some sort of DC clamp effect. However I then tried loading the output with 50R (as I didn't have 75R to hand) and the output fell to virtually nothing! The Rx is based on an AWI5822 IC, again it's datasheet revealed a 'typical' circuit which seemed identical to mine. I spent some time probing the surface mount output transistors, but since bias seemed OK DC wise I couldn't see any reason for the final emitter follower's lack of drive. Eventually the penny dropped, yet another of the ubiquitous 470uF caps was used to AC couple the video signal to the output socket. It was sited right next to the 7805 and its uF had obviously boiled away, replacement gave an excellent picture!

Now on the home run I modified the Tx to run on 5V rather than its 9V unregulated wallwart and fitted a 3V3 regulator where the 7805 had previously been. The Tx is now powered for 'free' by the Tivo USB port. The Rx remains wallwart powered, but since it will be turned off when not in use heat should not be too much of a problem.

The final picture shows the resulting image on the TV22.