Type tests and test houses
CPRI and ERDA type test reports, read properly
Three fields decide whether a certificate is worth anything to you: the sample tested, the class, and the date.
Two test houses, and what an outside report buys you
CPRI is the Central Power Research Institute, an autonomous body under the Government of India's Ministry of Power, with its principal laboratories at Bengaluru. ERDA is the Electrical Research & Development Association, a co-operative body of the electrical industry and the utilities, at Vadodara in Gujarat. They are the two names an Indian vendor file is most likely to put in front of you, and neither sells what you are buying.
That is the whole of their value. Most serious Indian makers already own the equipment: the factory behind this reference runs an in-house partial-discharge lab, a high-voltage lab and a general product-testing lab. The difference is not the kit in the bay. A works report is a self-declaration — the party with the commercial interest picked the sample and wrote the result. A national test-house report puts somebody else between the sample and the claim, and gives you a document number that lives outside the supplier's filing cabinet.
Neither is a licence to supply. It records that a sample of a stated design passed a stated sequence on a stated date.
One construction, one voltage class, one date
Read the scoping fields before the verdict. A report that reached a customer passed, so the result is the least informative line on the page.
Construction. A kit is an assembly, not a part number: a particular sleeve, stress-control component and mastic, over a particular conductor range, on a particular cable. Move from XLPE to paper-insulated, or three cores to four, and you have left the tested construction while the part-number family has not visibly moved.
Class. Voltage class does not round up, and a family sold across two classes does not merge their paperwork. Watch the notation. In Indian practice 11 kV and 33 kV are system voltages, while the 12 kV and 36 kV printed in test columns are the corresponding highest voltages for equipment. Mixed notations make a matched pair look mismatched, and the reverse. Take both numbers down.
Date. Two things age. The product ages, because a compound reformulation or a new moulding tool can put shipped stock outside the sample tested, and no certificate will say so. The standard ages too. A report written in 2012 followed the 2012 edition; ask which edition its schedule used and file the answer. The standards page covers what you should be citing in the first place.
33 kV is not 36 kV, and a class is not a range
The commonest way a genuine certificate ends up covering nothing is the class jump. On the file behind this reference the CPRI certificate covers 33 kV heat-shrink joints and terminations and is dated August 2012. If your requirement is 36 kV, that document does not reach it, however good the supplier is. Ask for the certificate naming your own class; if there is none, get that said in writing and price the gap.
How to check a report before it enters the tender file
This is this reference's own checking routine, not a procedure published by any test house or standards body. It is a document check, and it costs about ten minutes per certificate.
Ask for the number, not the picture
Report number, issuing laboratory and date of test, as three separate fields in the body of the email. A scanned front page proves nothing a clean PDF would not prove better.
Tick the sample description against your scope, line by line
Cable type, insulation, core count, conductor range, armouring, indoor or outdoor. One line you cannot tick disqualifies the document for this order, even though the report itself is perfectly valid.
Record the class in both notations
Copy the system voltage and the highest voltage for equipment exactly as printed. If only one is given, ask which it is. Never make that conversion yourself inside a bid evaluation.
Match the date to the edition you cited
Ask which edition and amendment of IEC 60502-4, CENELEC HD 629.1 or the relevant IS document the schedule followed. An older edition is not automatically a failure; an unanswered question is.
Ask which tests were actually run
A schedule is a sequence and a report may cover part of it. Ask for the tests performed with the result against each, not a covering line saying the sample passed.
Ask for the laboratory's accreditation scope
A test house is accredited for named tests within named ranges, not in general. Confirm the test you are relying on sits inside that scope at your voltage.
The shape of a published schedule: a 1.1 and 3.3 kV termination kit
| Test | Parameter for evaluation | 1.1 kV class | 3.3 kV class | Result |
|---|---|---|---|---|
| Insulation resistance | 1 min., 500 V DC | – | – | >10³ Ω |
| A.C. high voltage test (dry) | 15 min. | 4 | 8 | No breakdown & no flashover |
| Impulse withstand test at ambient temperature | 10 positive & 10 negative | 8 | 12 | No breakdown & no flashover |
| Heating cycle test | 63 cycles, 8-hour cycle, 5 hrs heating & 3 hrs cooling | – | – | No breakdown & no flashover |
| Insulation resistance after impact test (immersed) | 1 min., 500 V DC | – | – | >10³ Ω |
| Impulse withstand test at elevated temperature | 10 positive & 10 negative, conductor temperature 95 to 100 °C | 8 | 12 | No breakdown & no flashover |
| D.C. high voltage test (dry) | 5 min. | 15 | 21 | No breakdown & no flashover |
Transcribed from the GLT-1100 series kit data published by the manufacturer behind this reference, under the heading Guaranteed Technical Particulars in compliance with IEC 60502, 60060 & 60230. The two numeric columns are the rated classes the family covers; the figures beneath are the voltages applied. Count the columns before you read the numbers — a schedule usable in a bid names the test, what was evaluated, for how long, at what voltage against which class, and what counts as a pass. It is here as evidence structure, not product data: this site publishes no size charts, and the kit codes printed beside this schedule are deliberately left out. One figure needs flagging — the insulation-resistance limit appears as >10³ Ω in the catalogue and renders differently on the web page for the same kit, so confirm the exponent and the unit in writing.
Seven rows, and no partial-discharge measurement among them. That absence is correct: it is what a 1.1 and 3.3 kV schedule looks like. The medium-voltage schedule for the same maker's 12, 24 and 36 kV kits is written against IEC 60502-4 and CENELEC HD 629.1 instead, runs past twenty rows, and carries a 5 pC partial-discharge limit measured repeatedly — at ambient and again with the conductor at 95 to 100 °C — plus wet withstand, thermal short-circuit, immersion and salt-fog stages the low-voltage sequence has no use for. HD 629.1 sets test requirements for accessories on cables from 3.6/6 (7.2) kV up to 20.8/36 (42) kV, which is why that schedule stops where it does.
The use of knowing this is quick. If a 33 kV enquiry comes back answered with a seven-row low-voltage table, you can see the substitution from the row count without reading a figure.
Not published: the report numbers themselves
The manufacturer's own certifications page names a CPRI certificate for 33 kV heat-shrink joints and terminations dated August 2012, and an ERDA test report for the busbar sleeve. Neither carries a published report number, and the ERDA report is published with no date and no sample construction. Those are the fields step one above asks for, and they are not on the public record for either document. Ask for all of them in writing, per product, before the vendor file closes.
Two products, two reports, and why the pairing is the lesson
Read those two documents together rather than separately. The joints and terminations were certified by CPRI at 33 kV in August 2012. The busbar sleeve holds its own separate report from ERDA. One factory, one quality system — and still two test houses, two scopes, two dates, because they are two different products.
Certification does not attach to a company. It attaches to a construction, tested once, as a sample. A supplier holding both reports has proved two things and has not proved a third. So the enquiry question is not are you CPRI certified. It is which lines on this bill of materials carry an independent type test report, and which do not. Expect a mixed answer; be suspicious of a clean one that arrives without a report per line. The manufacturer behind this reference is Gala Thermo Shrink Pvt. Ltd., and what it holds is set out on the manufacturer page.
UL file numbers: what they prove and what they do not
Two UL files sit on this same vendor record — E328538 for the heat-shrink tubes, listed in 2009, and E335936 for the Busboot busbar insulating shroud and Bustube, listed in 2011. A file number is a lookup key, not a claim.
- A number is checkable; a logo is not. Search the file in UL's own directory and read back the listed products, the standard applied and the current status. A UL mark printed with no file number beside it is unverifiable by design.
- A listing covers the listed material, not the assembly. A file for heat-shrink tubing tells you about tubing. It says nothing about the kit that tubing goes into, and nothing about the installed joint.
- Check the status on the day you evaluate. Listings run under ongoing factory surveillance and can be withdrawn. A 2009 date says when the file opened, not that it is live now.
- A UL listing is not an IEC type test. Different standards, different questions. Offering one where the specification asked for the other belongs in the deviations schedule of the bid, not folded into a compliance column.
- Weight varies by market. In an Indian utility tender the IS or IEC type test is the requirement and the UL file is supporting evidence. Read the client's own approved-vendor rules rather than assuming which way it falls.
Sources disagree: KEMA
The manufacturer's joints and terminations catalogue prints a KEMA mark alongside the CPRI one, and a Tavanir approval letter in the same document set refers to both. The certifications page on the same website lists ISO, CPRI, ERDA, UL, MSEDCL, PGCIL, SEB and EDL, and does not mention KEMA. The sources do not agree and nothing published resolves it. If a KEMA report is load-bearing in your specification, ask for its number and the products it covers. A mark in a catalogue is not evidence of a report.
What to send when a certificate does not cover your order
Do not argue with the document and do not quietly file it. Send these three sentences, in this order, and keep the reply.
- The certificate you sent covers [construction] at [voltage class], dated [date]. Our requirement is [construction] at [voltage class]. Please confirm whether a type test report exists for our construction and our class.
- If one exists, send the report number, the issuing laboratory, the date of test, and the tests performed with the result recorded against each.
- If one does not exist, please confirm that in writing, and state which standard and edition the offered item is manufactured and routine tested to.
- The third sentence earns its place. A supplier who will write down that a report does not exist is one you can build a file around. A supplier who answers the first two and goes quiet on the third has answered it anyway — and that is the point to move the missing evidence into the acceptance tests on the purchase order.
Type tests are one layer. Approval letters are the next.
A test house records that a construction passed on a date. A utility approval admits a named supplier to a named voltage band for that utility's own projects, and it expires. Neither substitutes for the other.