Minikits 23cms RA18H1213G amp

[VK4UH]

Thanks for treading this

I have just completed assembly of a Minikits 23cms amplifier using the Mitsibushi RA18H1213G module which is supposed to produce about 20 watts PEP on 1296MHz

Construction has been uneventful. On set-up and testing, as per the instructions, I have set the bias voltage (VGG) on the module to 4.5 volts to find that the module wants to draw over 4 amps at 12 volts at this setting. The regulator circuit is working and the standing current can be reduced easily and smoothly.

The instructions however state that the bias voltage should not be run at less than 4.5 volts in SSB mode? There is no information given however on what level of standing current should be expected

I should have said that the amp has a 50R load on the input and a power meter and load on the output. There is no suggestion of any instability or oscillation.

I have not yet applied any RF drive.

My gut feeling is that 4 amps is excessive as standing current for such a small module without drive??

Has anyone else built one of these amps or have experience with these modules, who could tell me what value of standing current they observed.

Many thanks

Kevin Johnston VK4UH
Brisbane

[VK4ADM]

Hi Kevin
I am also using the RA18H1213G with my 23cm transverter,the standing current is 4A which peaks to 6a with drive .These chips are not that efficient on ssb and your standing current would be fairly typical.
Regards Daryl VK4ADM

[VK4GHZ]

KJ,

The module should be set to 4-5amps standing current*

The drive should be kept below 400mW to prevent damage to input FET.

P.S. from DK:
*There is some debate about the actual quiescent current setting.
I suggest setting the quiescent current to about 2.5A per RA18H initially for digital ATV use .
Increase it gradually towards 4A noting linearity or quality readings on digital tx.
Drive to a single chip will only need to be 150mW depending on desired output level.
Linearity drops off rapidly above 20W output per RA18H .
The above notes can also apply to other Mitsu chips, Often the cause of cutting out when chip is hot.. well worth checking the dc feeds for continuity before binning any chip.
Fitting small ferrite beads onto the legs can help reduce instability in the PA stage..

Source: http://www.g8ajn.tv/ra18h.html (Repair notes box)

4A seems typical?


I remember my 30W Spectrian 2.4 GHZ PA drew ~ 5A on idle.
An automotive relay was used to switch power on only during TX, a consideration if you're portable with these linear PAs and/or not running off mains.

[ZL2BKC]

Thanks for treading this

I have just completed assembly of a Minikits 23cms amplifier using the Mitsibushi RA18H1213G module which is supposed to produce about 20 watts PEP on 1296MHz

Kevin,

This is perfectly acceptable for the RA18H1213 power block. I have built several before and generally I like to set the idle bias current at 2A which according to my notes is Vgg=4.4V. However for the beacon at ZL2WHO/b more gain was required so the idle bias current is set to between 4 and 4.5A and that runs 24x7x365 with 18W out (Idd =~ 6A key down).

For DATV my final PA I have set the idle current closer to 8A for best linearity. If you check the datasheet there are several plots with 20W out and Idd=9A so we are still within spec. However this reduces the safety margin especially during bad SWR since the reflected power needs to be dissipated at the power block. So a good a ^%&^% good heatsink is required.

From the ones I have built I have found the gain to be higher at 1270MHz vs 1296 or 1250 MHz, so you do need a lot more drive to get the same power out. When the block is closer to Psat the gain levels out which is OK again for SSB, but has a negative impact for DATV.

The datasheet saying below 4.5V for SSB means it is starting to run class C at that point and no longer linear, so make the bias adjustable (which can be simple as a series resistor to the Vgg input which has a 2.5K input impedance) and set for a bias level of around 1 to 2A for SSB/CW work. You can set it to 5A if you need more gain due to a lower drive level - that is acceptable but you need to consider the heatsink and overall power consumption if used portable.

Personal preference is to make the Vgg supply adjustable +/- a volt or two and adjust to taste. Compared to the RA30 series for 2m and 70cm the 23cm block require an extra volt of Vgg before they start to bias.

Hope you make many contacts with the new setup.

73,
Wayne ZL2BKC

[VK4REX]

Hi Kevin

Have sent you an email about the mounting of the module, it is an application note from Mitsubishi, and explains the construction of the flange and how to properly mount it.

Rex

[VK4UH]

Thanks to everyone who offered help with this question.
Always good to see something working properly. Have tested now with RF drive and amp is working "just like it says on the box"
Thanks Also Rex for mounting info.

73 all
KJ VK4UH

[GM3SEK]

Hi Kevin

Have sent you an email about the mounting of the module, it is an application note from Mitsubishi, and explains the construction of the flange and how to properly mount it.

Rex

There's also some very good information from Down East Microwave: http://01895fa.netsolhost.com/PDF/MOSFET%20PA_pdf.PDF

Yes, 4A bias current looks OK. Remove bias on rx/standby, of course.

73 from Ian GM3SEK

[VK5PJ]

Hi,
the standing current is a lot higher that I had expected, was hopeful of a power efficient P.A for the 23cm beacon I am building but with 4amps all the of the key up time, is quite a current draw for idle over 24Hours ( I want to have the beacon run as a low duty cycle system if possible )

has ony one on the forum tried this module in Class C for CW or has all the experiences been in linear mode?

Regards,

[ZL2BKC]

Hi,
the standing current is a lot higher that I had expected, was hopeful of a power efficient P.A for the 23cm beacon I am building but with 4amps all the of the key up time, is quite a current draw for idle over 24Hours ( I want to have the beacon run as a low duty cycle system if possible )

has ony one on the forum tried this module in Class C for CW or has all the experiences been in linear mode?

Regards,

If it is low duty cycle I recommend you power off the bias supply as part of the CW keying sequence. This is what I do for the beacons I have built. In fact I have taken this one step further and placed a simple RC filter on the bias supply which performs envelope shaping to control the key clicks.

Using this approach means the current is almost zero when the carrier is off. As you have already found out the efficiency is not that great, but you don't get much choice at microwave frequencies. Even with low bias and extra drive the efficiency is still low and the only way to improve this would be to use a device with a higher Vdd

I'm happy to share more details here if interested. I have a PCB that includes a PHA-1 driver, opto-isolator, envelope shaping and LPF (for 6m and 2m) available as well. 3 minikits boards rolled into one with the same footprint as one.

73,
Wayne ZL2BKC