CUSTOM LAB EQUIPMENT WITHOUT THE INSTRUMENT-COMPANY MARKUP
Aknazar Arysbek, founder — biomedical engineer (MSc, Tufts University). Has worked across growing biotech startups and metal fabrication, on projects from custom bioreactors to high-pressure pharmaceutical extraction systems and custom medical equipment. LinkedIn ↗
Custom lab equipment gets built one of three ways: a commercial vendor quotes five figures, a machine shop builds the metal but not the electronics, or a graduate student wires a rig that leaves undocumented when they do. A fourth path, a small studio that does electronics, firmware, and mechanics as one documented job, costs a fraction and prices for a grant line. This note explains when to use each, and the under-$10,000 purchase-order lane a department can approve directly.
Every experimental lab eventually hits the same wall: the measurement or process the next paper depends on needs an instrument that does not exist in any catalog. A bioreactor with a specific geometry and sensing. A mechanical rig that loads samples the way the protocol demands. A jig that positions, heats, stirs, or logs something no vendor thought to build.
At that point most labs discover the market has a hole in it, and fall into one of three bad options.
The three bad options
The instrument giants. The big vendors are excellent at their catalogs and service contracts, but a one-off custom build is not their business. If they quote at all, the number starts in five figures, because their overhead structure allows nothing else.
The machine shop. A university shop does beautiful metal. But a modern rig is metal plus sensors, actuators, a control board, and firmware, and the shop's lathe does not write firmware. You become the systems integrator of your own instrument.
The grad student. A capable student wires a microcontroller to a pile of parts, it works, papers get published, then the student graduates. The lab inherits an undocumented rig nobody dares to touch. Every PI reading this has seen one, and many are running one now.
The fourth option barely exists: a small engineering studio that does electronics, firmware, and mechanics as one job, documents everything, and prices for a grant line rather than a corporate procurement budget. Almost no freelancers take these jobs, because each one needs three disciplines at once. That gap is precisely where we work.
Why custom quotes come back at five figures
Fairness requires saying that the big quotes are not usually gouging. A custom instrument from a commercial vendor carries real costs: engineering hours across several disciplines, one-off design work amortized over exactly one unit, liability and warranty overhead, calibration traceability, and the sales machinery around it. If you need certified, validated, GxP-grade equipment, pay for that, because it is worth it.
But most research rigs do not need any of that. They need to work, be understood, and be repairable. When a lab pays instrument-company overhead for a research-grade one-off, the extra money buys the vendor's cost structure, not better science. A leaner builder with the right skills delivers the same function for a fraction, provided the build is documented properly.
Catalog, DIY, or custom: a 30-second test
| Situation | Right answer |
|---|---|
| A catalog instrument does 90% of what you need | Buy it. Adapt the protocol. Custom exists for the cases where adaptation breaks the science. |
| A student wants to build it, and it is scientifically interesting to build | Let them. It is training. But require documentation as a deliverable, not an afterthought. |
| The rig is critical to the lab's output for years, and nobody's career should depend on maintaining it | Have it built professionally, with docs, a spares list, and source code in the lab's hands. |
| You need certified or validated equipment for regulated work | Commercial vendor. That overhead is the product. |
What a custom rig should ship with
The difference between an asset and a liability is the documentation. Whatever you commission, from anyone, require the following.
- Schematics and CAD source, the editable files, not PDFs of them, in your hands.
- Firmware source code, commented, with build instructions a new student can follow.
- A bill of materials with real part numbers, so a failed sensor is a $30 order, not a service call.
- An operating and maintenance writeup thorough enough to cite in a Methods section, because reviewers increasingly ask and replication is the point.
- A calibration procedure, if the rig measures anything you will publish.
We build open by conviction; our own products are open-source hardware, so we hand all of this over as a matter of course.
The procurement path: use the small-purchase lane
Buying a custom rig can be faster than buying a big instrument. Most US universities let a department issue a purchase order directly for purchases under a threshold, commonly $10,000, with no competitive bidding. Between roughly $10,000 and $50,000 you typically need two or three written quotes, but still no formal RFP. The mechanics that make it smooth are as follows.
A fixed, itemized quote drops straight into a PO request or a grant budget justification. Engineering goes on one line, materials at documented cost. Grants routinely allow an equipment line to include delivery, installation, and calibration.
Phase the work to fit the lane. A feasibility study, which we price at $1,500 to $3,000, is a small purchase anywhere. Its deliverable, a concept design with parts cost and a fixed build quote, is exactly the document a grant budget or quote comparison needs for the build itself.
Vendor registration is routine. We are a US LLC and can register in university supplier portals and invoice against POs.
What it costs with us
Feasibility brief: $1,500 to $3,000, about two weeks, covering whether it can be built, how, concept CAD, and parts cost. Working instrument: $8,000 to $25,000 depending on complexity, materials at cost with receipts. Prices are fixed before work starts, and the build is measured against a written acceptance checklist; if it does not pass, we keep working at no extra charge. The founder is a biomedical engineer who has built programmable bioreactors and lab robots, so your rig is designed by someone who has stood at the bench. Full pricing and guarantees are on the pricing page.
Common questions
Who can build custom lab equipment for a research lab? Three sources exist today: commercial instrument vendors (five figures, certified, slow), a university machine shop (mechanics only, no electronics or firmware), and a graduate student (cheap, undocumented, leaves when they graduate). A small engineering studio that does electronics, firmware, and mechanics together, with documentation, fills the gap between them and prices for a grant line.
How much does a custom bioreactor or test rig cost to design and build? A feasibility study runs $1,500 to $3,000. A working custom instrument runs $8,000 to $25,000 with a lean studio, materials at cost. A commercial vendor's one-off starts in five figures because its overhead structure requires it, which is worth paying only when you need certified or validated equipment.
Can a university buy custom equipment without competitive bidding? Usually yes under a dollar threshold, commonly $10,000, where a department issues a purchase order directly. Between roughly $10,000 and $50,000 you typically need two or three written quotes but no formal RFP. A fixed, itemized quote fits either lane and a grant equipment budget.
What should a custom lab instrument include besides the hardware? Editable schematics and CAD source, commented firmware source with build instructions, a bill of materials with real part numbers, an operating and maintenance writeup detailed enough to cite in a Methods section, and a calibration procedure if it measures anything you will publish. Without these the rig becomes a liability the day its builder leaves.
Get a quote-ready feasibility brief →
Procurement references: VCU: procurement methods & thresholds · BidFinderEDU: how university procurement works · MIT RAS: equipment in grant budgets. Founders & owners: see the companion note on what a prototype actually costs.