Is the CGH40006P Still in Production?

Yes. The transistor is active; the boards that carry it are the retired items. Distributor records list the CGH40006P itself as Active, with a 26-week factory lead time.

The CGH40006P-AMP demonstration amplifier is listed as not recommended for new designs, and the CGH40006P-TB demo board is discontinued. When a comparison page reads one of those board statuses and prints it beside the device part number, a live part looks obsolete.

The current datasheet is Rev 3.3, dated 2022-11-01, headed MACOM Technology Solutions Inc.

What Can Replace the CGH40006P?

Three same-family options, and none of them shares the 440109 footprint. Each solves a different problem, and each changes something the original did not.

Read the package-type row first. Four parts, four different numbers, four separate outline drawings. Not one of these is a drop-in replacement for the CGH40006P - and the CGH40006S, the closest in power, is the worst offender because its junction rating is 50 °C lower.

The 10 W pill sits two steps up the same 28 V, DC–6 GHz ladder - and it is the one worth looking at first, because it keeps the original's junction rating.

The 10 W Pill: Same Three Terminals, Different Footprint

Same pin functions, same junction limit, different outline and a different dimension scheme. The CGH40010P is the closest thing to a drop-in the family offers - and it still is not one.

What carries over: the terminal assignment. Both pills bring out pin 1 GATE, pin 2 DRAIN, pin 3 SOURCE in that order. Both are rated to the same 225 °C junction. Both run on the same 28 V rail over the same DC–6 GHz band.

What does not: the outline. The CGH40006P uses package type 440109; the CGH40010P uses 440196. Each has its own drawing.

The two drawings do not even share a dimension lettering scheme. The letter G is a few-thousandths detail on the 440109, and the 13.84 to 14.09 mm overall length on the 440196.

Both pills carry the same three terminal functions in the same order. That is the only thing you can carry across - the outline drawings are separate documents, and the land pattern, stencil and thermal pad all change with them.

What Else Moves When You Go to 10 W?

Current capability doubles, case resistance drops, and the datasheet changes the frequency it characterises at. Three of those four changes work in your favour, and one of them makes the numbers harder to compare.

The last row is the one to watch. The 6 W part is characterised at 2.0 GHz and the 10 W part at 3.7 GHz - different operating points, so their gain and power figures cannot be read across from one datasheet to the other.

And one of these numbers carries a hidden condition

The 8.0 °C/W figure comes with a footnote that says it was measured for the CGH40010F - the screw-down flange version - at 14 W dissipation. The 6 W part's 9.5 °C/W was measured for the P, at 8 W.

So the two case-resistance numbers describe different package variants at different dissipation levels. The direction of the improvement is real; the exact ratio is not something you can read off the two datasheets.

Scaled within each pair. The junction rating is the reason to consider the 10 W part at all - it is the one thing the QFN alternative gives away.

Migration Checklist: 6 W Pill to 10 W Pill

Treat it as a re-layout with a bias-point change, not a substitution. Six items to work through before you commit the board.

1. Take the land pattern from the 440196 drawing. Do not scale or reuse the 440109 footprint, and do not try to map dimensions letter by letter - the two drawings label them differently.
2. Move the bias point. Quiescent current goes from 100 mA to 200 mA, and the gate quiescent voltage is quoted at that new point.
3. Re-check the gate drive budget. Maximum forward gate current doubles to 4.0 mA. If your bias network was sized tight for 2.1 mA, re-size it.
4. Re-run the input and output match. Input capacitance rises from 3.0 pF to 4.5 pF and output from 1.1 pF to 1.3 pF. A match tuned in the 6 W package will not transfer.
5. Re-quote the RF numbers from the 3.7 GHz table - or from data taken at the frequency you actually use. Do not scale the 2.0 GHz figures.
6. Re-do the thermal budget at the new dissipation. Running a 10 W device where a 6 W device sat means more heat, even though the case resistance figure is lower.

When to Stay With the Original, and When to Substitute

Stay with the CGH40006P when:

• Your board is already laid out for the 440109 pill and the design meets spec. The part is active, and a re-layout is the expensive option.
• Your output requirement sits at 6 W. Moving to a 10 W part costs you a re-layout and a re-match for headroom you may not need.
• You need the widest possible documentation set - the 6 W part has the AMP reference amplifier, S-parameter files and application notes alongside it.

Move to a substitute when:

• You need more output power on the same 28 V rail and the same 225 °C junction limit. That is the CGH40010P case, and it is the strongest one.
• Your layout can take a 3 mm × 3 mm surface-mount part and you can live with a 175 °C junction limit - the CGH40006S is the smaller board option, but it gives away 50 °C of thermal window.
• You are building your own package or module and want the die. The CGH60008D is an 8 W die rated to the same 225 °C junction, supplied in Gel-Pak containers.
• The 440109 footprint is the thing blocking you. If you must re-layout anyway, re-layout onto the part you actually want.

Where to Buy CGH40006P Replacements

We stock the CGH40006P and can quote the CGH40006S, CGH40010P and CGH60008D against the same requirement. When you send us a target output power and your board constraints, we will tell you which of the four actually fits - and which one forces a re-layout.

Tell us the frequency, the drive level and the junction temperature you are designing to, and we will confirm the part and quote on your quantity.

Frequently Asked Questions About CGH40006P Replacements

Q1: Is there a pin-compatible replacement for the CGH40006P?

A: No. The three closest same-family alternatives use package types 440203, 440196 and a bare die respectively, against the CGH40006P's 440109. Each has its own outline drawing. Same family, no shared footprint.

Q2: Is the CGH40006S a drop-in for the CGH40006P?

A: No, and it is the riskiest of the three to assume. It is a 3 mm × 3 mm QFN with six pads against the pill's three leads, and its junction rating is 175 °C against the pill's 225 °C.

Q3: Which replacement keeps the 225 °C junction rating?

A: The CGH40010P and the CGH60008D. Both are rated to 225 °C, the same as the CGH40006P. The CGH40006S drops to 175 °C.

Q4: Can I use the CGH40010P in place of the CGH40006P without a board change?

A: No. The terminals match - pin 1 gate, pin 2 drain, pin 3 source on both - but the package type differs (440196 against 440109), and the two outline drawings do not even share a dimension lettering scheme.

Q5: What changes electrically when I move up to 10 W?

A: Four things. Max drain current doubles from 0.75 A to 1.5 A, quiescent bias moves from 100 mA to 200 mA, input capacitance rises from 3.0 pF to 4.5 pF, and the datasheet characterises the RF performance at 3.7 GHz instead of 2.0 GHz.

Q6: Is the 8.0 °C/W figure on the CGH40010 directly comparable to the 9.5 °C/W on the CGH40006P?

A: Not exactly. The 8.0 °C/W is footnoted as measured for the CGH40010F, the screw-down flange version, at 14 W dissipation. The 6 W figure was measured for the P variant at 8 W.

Q7: What is the CGH60008D?

A: An 8 W GaN HEMT supplied as a bare die, rated to a 225 °C junction and the same 0.75 A drain current as the CGH40006P, with 15 dB typical gain at 4 GHz. It ships in Gel-Pak containers and is aimed at builders who make their own package or module.

Q8: Are the higher-power family members - 45 W, 90 W, 120 W, 180 W - replacements?

A: They are the same family, not replacements. They run on the same 28 V rail over the same band, but the power step is far larger than a substitution implies, and each has its own package and layout.

Q9: Is the CGH40006P itself being discontinued?

A: No. The transistor is active with a 26-week factory lead time. What has been retired is the support hardware: the CGH40006P-AMP demonstration amplifier is not recommended for new designs, and the CGH40006P-TB demo board is discontinued.

Q10: How do I compare the gain figures across these parts?

A: Don't - not directly. Each datasheet quotes its RF table at a different frequency: 2.0 GHz for the CGH40006P, 5.8 GHz for the CGH40006S, 3.7 GHz for the CGH40010P, and 4.0 / 6.0 GHz for the CGH60008D. Pull each part's figure at your own operating frequency.