Solid State amplifier problems

[Kennrth]

This is a warning to all hams who want to use solid state linear amplifiers with antenna tuners. It hurts the people who can afford it the least. Hams that use long wire antennas, G5RV’s any antennas that hams’ use antenna tuners for to work on multiple bands.

LDMOS solid state amplifier devices are capable of amplifying frequencies up to 500mhz. Linear amplifiers unlike for example amplifiers in your HF radio have no feedback controls. Any circuit that has more than 3db gain is subject to oscillation. There is both dc ad ac feedback methods to control the stability of the amp.

There is on the sign wave of solid state linear amplifiers a fuzz that rides on the sine wave at very low level. Its more apparent in amps that use two or more LDMOS devices. The signal source from your drive may be clean (HF Radio) but parasitic oscillations are created many times by an imbalance of gain between amplifier devices (LDMOS). Also from capacitance s and inductance can create very low level parasitic. Ceramic capacitance is notorious for this because they are microphonic. Many linear amplifiers are designed to amplifier 160m to 10m bands. 1500Khz to 30mhz. Some can work on the 50mhz band.

These ultra high frequency parasitics cannot flow down your transmission line let alone be radiated by you antenna. Every time you amplifier sends big power down the line. The antenna tuner passes the parasitic energy on the air tuners capacitor plates and and inductor storing the high frequency parasitic energy that reflects back in the LDMOS gates. Causing out of band oscillation but the LDMOS has no problem amplifying going into high energy out of band oscillation blowing the gates of the LDMOS devices. There is no feedback control or swamping mechanism. Also high input impedance between the tuner and the amp. Cause spikes on the transmission line by standing wave in phase with the output sine wave. The high voltage spikes also is a failure mode for LDMOS devices. LSMOS devices are limited to 50V by design on the source and the gate is not supposed not to pass dc current. Large spikes on the gate specially in reverse will blow them up.

Engineers are running these devices full out and require extraordinary methods of sinking heat. In most cases the LDMOS devices are not failing from overheating but uncontrolled oscillation that happens in milliseconds or less.

If you are using an antenna tuner to match the high impedance of your antenna do not use a solid state linear. Buy a tube linear. A 3x500Z or ZG type tube is preferable.

Sad that Ameritron is out business. They made several models that used glass 3x500 tubes. Ameritron AL 811 is also a good choice. Many used ones available now since many people looking to move up to legal limit. Going from 1KW to 1.5KW will only make .6 increase on the receivers S meter. You have to quadruple you output to make 1 S unit change on the receivers radio. When the band is open 100W is usually sufficient. They are much more rugged than very expensive solid state linears. Acom and OM power make nice tube amps just have to be careful tuning and driving them. They are much more forgiving than solid state amps.

There are linear amps that use other types of tubes many metal type with very high gain. They are often blown up by over driving them by accident. A 50W will drive1500W + output. And accidents in tuning properly your antenna tuner before full power output is achieved.

HIgh Power LDMOS devices are around $250 a piece. Unlike tubes they are very difficult to replace.
Very difficult replacing them in Expert Linear amp company. Easier on Mercury amps but still difficult requires very large solder irons etc. And skills.

 

[RFI-EMI-GUY]

If these oscillations are out of band, and any reactive device after the amplifier sustains these, for those with these amplifiers, what is the practical solution?

 

[prcguy]

In my experience, all the potential problems listed above can be addressed with a good design. I have high power LDMOS amps that use a diplexer instead of a low pass filter on the output. All reflected harmonics are absorbed by an onboard load and the LDMOS never sees it.

My amp also has excellent protection from overload using two types of diodes across the gates to hard clamp any RF above the 2 watt drive level spec of the 1200w output LDMOS. I use various onboard attenuators to set the input level of this particular amp to either 10w or 40w input depending on what radio I'm driving it with. I've accidently hit the amp with 100w when set up for 10w and did blow the input protection diodes but the failure mode of the diodes is a short which further protects the LDMOS. For a few $ I replaced the input protection diodes and back in business.

There is also SWR protection that senses anything over 2:1 and if you exceed that both the 50v rail and bias are immediately removed protecting the LDMOS. The power/SWR meter has a microprocessor so it will measure SWR at any power level and can shut down the amp due to high SWR with only a few watts when you first key up.

Some mfrs, especially the CB amp guys who have discovered LDMOS are building them cheap with little or no protection and they will be much more susceptible to LDMOS failure.

I've had quite a few tube amps and many that used 3-500Zs and have had lots of problems with them like high power parasitic oscillations that destroyed tubes and power supplies. And tubes are getting really expensive. A good LDMOS amp can run a lifetime with no failures and nothing to replace.

 

[RFI-EMI-GUY]

Some mfrs, especially the CB amp guys who have discovered LDMOS are building them cheap with little or no protection and they will be much more susceptible to LDMOS failure.

Yeah I have seen some videos where homemade LDMOS CB amps have simply some unshielded hookup wire between the T/R relay (open frame standard contactor) and the input and output SO239's.

 

[prcguy]

Yeah I have seen some videos where homemade LDMOS CB amps have simply some unshielded hookup wire between the T/R relay (open frame standard contactor) and the input and output SO239's.

And virtually no RF bypass on the B+ lead into the amp to the LDMOS drain. The entire inside of the amp and power leads are hot with RF.

 

[Kennrth]

In my experience, all the potential problems listed above can be addressed with a good design. I have high power LDMOS amps that use a diplexer instead of a low pass filter on the output. All reflected harmonics are absorbed by an onboard load and the LDMOS never sees it.

My amp also has excellent protection from overload using two types of diodes across the gates to hard clamp any RF above the 2 watt drive level spec of the 1200w output LDMOS. I use various onboard attenuators to set the input level of this particular amp to either 10w or 40w input depending on what radio I'm driving it with. I've accidently hit the amp with 100w when set up for 10w and did blow the input protection diodes but the failure mode of the diodes is a short which further protects the LDMOS. For a few $ I replaced the input protection diodes and back in business.

There is also SWR protection that senses anything over 2:1 and if you exceed that both the 50v rail and bias are immediately removed protecting the LDMOS. The power/SWR meter has a microprocessor so it will measure SWR at any power level and can shut down the amp due to high SWR with only a few watts when you first key up.

Some mfrs, especially the CB amp guys who have discovered LDMOS are building them cheap with little or no protection and they will be much more susceptible to LDMOS failure.

I've had quite a few tube amps and many that used 3-500Zs and have had lots of problems with them like high power parasitic oscillations that destroyed tubes and power supplies. And tubes are getting really expensive. A good LDMOS amp can run a lifetime with no failures and nothing to replace.

Agree .
Yes, 3-500Z power triodes are highly prone to VHF (Very High Frequency) parasitic oscillations. Because of this, amplifiers using these tubes—such as the Ameritron AL-80B or Heathkit SB-220—rely on Parasitic Suppressors (a small coil and resistor in parallel) on the anode to stabilize them. [1, 2, 3, 4, 5]

Why Parasitic s Happen

• Internal Resonance: The tube's physical internal grid-to-cathode and plate structures self-resonate, usually in the 95 to 150 MHz range. [1, 2]
• Destructive Effects: If left unsuppressed, these oscillations can cause massive grid current spikes, arcing on bandswitches or tuning capacitors, and may even cause the hot tungsten filaments to physically bend and short. [1]

The Essential Fix

• Suppressors: You must use proper parasitic suppressors in the plate circuit to dampen the VHF resonance and lower its Q factor.
• Grounding: Control grids should be grounded directly to the chassis with nearly zero-length connections to keep the grid firmly at ground potential at high frequencies. [1, 2, 3, 4]

If you are experiencing issues like intermittent arcing or blown resistors in your 3-500Z amplifier, the parasitic suppressor is the first component you should inspect. [1, 2]

Note in Mercury Lux user manual states not to use their amplifier with an antenna tuner.
Expert Linear manual - didn't see a warning. The other day on 80 meters a pal of mine blew up his expert 1.5 K linear for the second time using antenna tuner, he did not make any tuning errors. There are many complaints about solid state amp going boom.



8 sites



Show all

 

[Kennrth]

If these oscillations are out of band, and any reactive device after the amplifier sustains these, for those with these amplifiers, what is the practical solution?

Avoid using solid state linear amps with antenna tunners

 

[prcguy]

Avoid using solid state linear amps with antenna tunners

Depends on the amp and some big LDMOS amps come with a built in auto antenna tuner. I use an LDG-1000 pro with my LDMOS amp depending on antenna but I never hit the tune button when the amp is on. I have a couple of old SGC 500w amps and have always used them with a LDG-600w tuner and never had a problem.

 

[Kennrth]

Agree .
Yes, 3-500Z power triodes are highly prone to VHF (Very High Frequency) parasitic oscillations. Because of this, amplifiers using these tubes—such as the Ameritron AL-80B or Heathkit SB-220—rely on Parasitic Suppressors (a small coil and resistor in parallel) on the anode to stabilize them. [1, 2, 3, 4, 5]

Why Parasitic s Happen

• Internal Resonance: The tube's physical internal grid-to-cathode and plate structures self-resonate, usually in the 95 to 150 MHz range. [1, 2]
• Destructive Effects: If left unsuppressed, these oscillations can cause massive grid current spikes, arcing on bandswitches or tuning capacitors, and may even cause the hot tungsten filaments to physically bend and short. [1]

The Essential Fix

• Suppressors: You must use proper parasitic suppressors in the plate circuit to dampen the VHF resonance and lower its Q factor.
• Grounding: Control grids should be grounded directly to the chassis with nearly zero-length connections to keep the grid firmly at ground potential at high frequencies. [1, 2, 3, 4]

If you are experiencing issues like intermittent arcing or blown resistors in your 3-500Z amplifier, the parasitic suppressor is the first component you should inspect. [1, 2]

Note in Mercury Lux user manual states not to use their amplifier with an antenna tuner.
Expert Linear manual - didn't see a warning. The other day on 80 meters a pal of mine blew up his expert 1.5 K linear for the second time using antenna tuner, he did not make any tuning errors. There are many complaints about solid state amp going boom.


View attachment 202437View attachment 202438View attachment 202431
8 sites



Show all

I worry also that the LDMOS parts maybe mishandled in transit or during installation. They are static prone and can have latent failures due to small static punches in semiconductor. LDMOS (Laterally Diffused Metal Oxide Semiconductor) devices are highly susceptible to damage from electrostatic discharge (ESD), commonly known as static.

 

[Kennrth]

See https://www.eham.net/reviews/view-product/12141?page=5&per-page=10
main complaint LDMOS blowing up sometimes multiple times .

 

[prcguy]

I worry also that the LDMOS parts maybe mishandled in transit or during installation. They are static prone and can have latent failures due to small static punches in semiconductor. LDMOS (Laterally Diffused Metal Oxide Semiconductor) devices are highly susceptible to damage from electrostatic discharge (ESD), commonly known as static.

Yes, I'm aware of all these things and have years of training handling static sensitive parts. Most commercial mfrs of LDMOS amps will have good anti static procedures but home builders and CB amp shops usually not.

 

[Kennrth]

I worked in a company for 30 years that manufactured both commercial and military equipment. I have seen numerous ESD violations. Even with quality assurance over sight. Starting with vendors shipping static sensitive parts in the incorrect packaging. Violations in receiving, expeditors , stock room, manufacturing floor , inspection, test department and management putting their hands where they don’t belong. We had a quality assurance department that over sees all of this. They find violations on ESD benches people wiping down antistatic mats with incorrect solution. Broken grounds and wrong resistors between mats and earth ground. Incorrect clothing and shoes. Broke wrist strap wires and wrist straps not properly being worn. Mats that failed antistatic tests. QA has had to terminate people who keep violating the procedures. You may be shocked as to how many people who think they are doing the right thing especially when there is no quality assurance oversight and regular esd bench testing.

I disagree with your assessment on solid state liner amplifiers. They are blowing up far more often than new tube failures. Many of these companies are pulling their hair out dealing with this especially under warrantee always blaming the customer.

Out of band parasitic oscillations of LDMOS devices are above 450 Mhz sometimes as high as 2 or 3 ghz. You will not see this on the oscilloscopes these shops have. Over weeks sometimes days it can destroy a LDMOS. Capacitors being used in these amps are not microwave rated parts.

And LDMOS amp will not pass a scratch test. Tube amps do without blowing the tube.

A scratch test involves a file or rasp connected to the amp output. Usually mounted in a safe box where the file is safely isolated. Then a heavy ground wire with a bare end is installed. Using a welder’s mask and heavy insulated gloves the tester scratches the file with the bare end of the wire for 10 seconds or more. This is done when transmitting a CW wave at full power. It is repeated usually In three parts of the freq output low med high freq. Bypassing the internal swr shut down circuitry.

Second part of the test involves high frequency band scratch same as above but a typical HV line filter is used between the file and the amp.

.

A good tube amp may blow a fuse shutting down even may damage some components, door knob capacitor for one but the tube will survive. At that point test engineering must assess any failure and find corrections. I for one know LDMOS devices will not survive this test. They stopped doing these tests because they could not figure out away to protect the LDMOS devices.

I strongly advise hams not to use solid state amps with antennas that need tuners.

Mercury Linear Amp company has warned their customers as well.

You do what you want its your property.

I also strongly disagree the statement that semi conductor last forever. They fail all the time.

They are subject to damage from high field pulses EMP, and cosmic radiation. There are times when the earth is subject a bursts of strong cosmic radiation. Satellite companies are deathly afraid of this. So are power companies.



Semiconductors degrade over time primarily due to atomic-level physical wear and environmental stress. As millions of microscopic transistors and interconnects operate, the continuous flow of electrical current, temperature shifts, and chemical interactions slowly alter their structural integrity, eventually leading to device failure



 Electromigration: The continuous flow of electric current physically pushes atoms in the microscopic metal pathways (interconnects). Over time, this creates tiny voids (leading to open circuits) or piles up material (causing short circuits). [1]

 Thermal Fatigue & Stress: Heating and cooling cycles cause the different materials in a semiconductor (silicon, copper, plastic packaging) to expand and contract at different rates. This repeated stress creates microscopic cracks, leading to delamination or broken connections. [1]

 Time-Dependent Dielectric Breakdown (TDDB): The insulating layers inside a transistor (such as the gate oxide) are incredibly thin. Under constant voltage stress, defects build up within this layer until a conductive path forms, causing an irreversible electrical short.



 Hot Carrier Injection (HCI): High-energy electrons can get injected into the insulating layers, becoming trapped. This shifts the electrical properties of the transistor over time, causing it to switch slower or require more power to operate.

 Mobile Ion Contamination: Tiny amounts of impurities (such as sodium or potassium ions) can diffuse into the semiconductor structure when exposed to heat and moisture. These roaming ions disrupt the intended electrical flow, causing performance drift.

PS I said nothing regarding Garage builders or CB amp builders.

The military has used this scratch testing their most critical products.

PS I said nothing regarding Garage builders or CB amp builders.

 

[Kennrth]

I just want to mention scratch testing is done with SWR protection disabled. A destructive test assess worst case conditions in first article testing. In production open and shorted output tests are done as well as scratch tests to each amp to making sure the protection circuitry is correctly built and installed. Before the amp leaves the factory. I for one don't believe most amateur radio amp builders are doing this

 

[prcguy]

I worked in Aerospace manufacturing and our anti static training, procedures and testing were top notch. I worked in assembly for both ground and airborne electronics and also did some time in a Gallium Arsenide foundry class 100 clean room, full bunny suits, extreme sensitivity to static, etc. I did everything from sawing out hundreds of tiny 10-40GHz LNAs from 3" GaAs wafers, wire bonding the LNAs into receivers and testing the LNAs and receivers with zero failures attributed to static. I suspect most serious companies have a similar track record.

A major problem in the LDMOS amplifier industry is Chinese fake transistors that get into the supply chain and will fail 100%. That has been a major problem in the hammy HF amplifier circles using LDMOS and even silicon transistors these days.

On my various tube amps, my Kenwood TL-922A would fart blue and purple lightning across many of the output components from high power VHF oscillations in the 3-500Zs. I had to double up the VHF snubbers on the plate caps which reduced power on 10m and I eventually dumped that amp. The new owner swapped out the tubes and it behaved better and he didn't use 10m. I had two Henry 2K-4s with 3-500Zs and something happened to one tube in each amp taking out the diode stack and HV choke in one and just the HV diodes in the other. I dumped those not being able to find a cheap enough replacement HV choke. I had a Collins 30S-1 that blew up real bad and I took it to the most respected amp guy in So Cal. He had it for about 2yrs and could not fix it. I've had several other big tube amps and none that I remember worked reliably.

My 1.2kW LDMOS amp kit I built has been flawless and lived through a number of screw ups I did that should have taken out the BLF188 transistor but its still going.

 

[Kennrth]

Life is good when every thing works